Best MCP Servers

claude-code-best-practice

shanraisshan/claude-code-best-practice · ★ 66k ▲149/7d · HTML

This is a comprehensive collection of best practices, guides, and configuration examples for Claude Code, an advanced AI coding assistant. It serves as a community-driven manual to help developers master the tool's powerful features, from basic settings to complex 'agentic engineering' workflows.

Verdict This repository is an indispensable, S-tier resource for any serious Claude Code user. It's the missing manual that expertly bridges the gap between official documentation and real-world mastery.

goose

aaif-goose/goose · ★ 54k ▲284/7d · Rust

Goose is a versatile, open-source AI agent that runs locally on your machine. It provides a unified interface, as a desktop app, a command-line tool, or an API, to interact with over 15 different AI model providers, from OpenAI to local models via Ollama. It aims to be the central hub for all your AI-powered tasks, including coding, research, and workflow automation.

Verdict Goose is an exceptional tool that delivers on its promise of being a universal AI agent. Its multi-platform support (desktop, CLI, API) and vast provider compatibility make it a powerful central hub for anyone who uses AI seriously.

repomix

yamadashy/repomix · ★ 28k ▲112/7d · TypeScript

Repomix gathers a code repository into one structured file that an AI assistant can read without opening every source file separately. It handles filtering, token counting, remote repositories, and smaller views of large codebases through a command-line tool, web app, library, MCP server, and Claude Code plugins.

Verdict Our Repomix run installed 371 packages in 22 seconds, passed all 1,813 tests, and found 0 known vulnerabilities, making it a low-friction choice for developers who need repository context for AI tools. Use it for local packing and controlled agent workflows, especially when filters and token budgets matter.

opencode

sst/opencode · ★ 208k ▲1.7k/7d · TypeScript

OpenCode is an open-source coding agent that runs in a terminal, desktop app, IDE extension, or server-backed web interface. It lets a language model inspect and edit a repository, run commands, call subagents, and use local or remote MCP tools under configurable permissions.

Verdict OpenCode is one of the strongest open choices for developers who want provider freedom without giving up a polished terminal agent, desktop client, or MCP support. Its permission system is capable, but the permissive defaults deserve deliberate configuration, and contributing to the giant monorepo is far heavier than installing the app.

learn-claude-code

shareAI-lab/learn-claude-code · ★ 77k ▲446/7d · Python

Learn Claude Code is a 17-chapter Python course that builds a small coding-agent harness one mechanism at a time. It teaches the loop around a model, including tools, permissions, hooks, planning, subagents, skills, memory, background work, teams, MCP, saved workflows, and goal-controlled stopping; English is the default, with Chinese and Japanese translations.

Verdict Our learn-claude-code run installed 48 packages in 24 seconds, built in 9 seconds, and passed its available tests in 31 seconds with 0 known audited vulnerabilities. It is a strong course for engineers who want to see an agent loop grow into permissions, memory, teams, MCP, workflows, and goal checks without starting inside a large framework.

kong

Kong/kong · ★ 44k ▲27/7d · Lua

Kong Gateway sits between clients and APIs, LLM providers, or MCP servers to route traffic and apply shared controls. It handles load balancing, health checks, authentication, rate limits, TLS, logging, transformations, multi-provider AI requests, and MCP traffic policy through configuration and plugins.

Verdict Our Kong proxy-Wasm fixture built in 22 seconds and passed all 6 tests, but that result covers only a small Rust directory inside the gateway repository. Kong is a sensible shortlist choice when one boundary must govern APIs, several LLM providers, and MCP traffic.

github-mcp-server

github/github-mcp-server · ★ 33k ▲143/7d · Go

GitHub MCP Server is GitHub's official bridge between Model Context Protocol clients and GitHub APIs. It lets coding agents read repositories, inspect issues and pull requests, query Actions and security data, and perform authorized changes through structured tools instead of screen scraping or improvised shell commands.

Verdict Our GitHub MCP Server run built in 42 seconds and passed all 29 tests in 28 seconds, giving the official integration a clean measured baseline. It is the first GitHub MCP server to try, provided you start with read-only mode, narrow toolsets, and the smallest useful OAuth scopes.

crush

charmbracelet/crush · ★ 28k ▲160/7d · Go

Crush is a terminal coding agent that can inspect and edit a project, run tools with permission, use language servers, and connect to several model providers. It keeps work in sessions and extends its tool set through MCP servers and Agent Skills.

Verdict Our Crush run built in 155 seconds and passed all 52 tests in 192 seconds, making this the cleanest contributor result in this batch so far. Use it if you want a polished terminal agent with real provider choice, MCP support, skills, and permission controls.

fastmcp

PrefectHQ/fastmcp · ★ 28k ▲116/7d · Python

FastMCP is a Python framework for building Model Context Protocol servers, clients, and interactive MCP apps. It turns ordinary functions into tools with generated schemas and handles protocol details such as transports, validation, lifecycle, and authentication.

Verdict Our FastMCP run installed 95 packages and built in 9 seconds, but test collection failed after 9 seconds on a missing OpenTelemetry SDK module. Use FastMCP when a Python team wants productive server and client abstractions and will test authentication as a separate subsystem.

Awesome-MCP-ZH

yzfly/Awesome-MCP-ZH · ★ 7.7k ▲10/7d

This is a comprehensive, curated directory of resources for the Model Context Protocol (MCP), an open standard that lets AI models interact with external tools and data. It's built specifically for a Chinese-speaking audience, collecting everything from beginner explanations and client software recommendations to a massive, categorized list of MCP servers that give AI new capabilities. It solves the discovery problem for anyone in the Chinese tech community looking to build with or use AI agents.

Verdict For any Mandarin speaker interested in the AI agent ecosystem, Awesome-MCP-ZH is an indispensable and brilliantly executed resource. It's more than a list; it's a curated gateway that provides context, practical advice, and a comprehensive map of a complex new field.

awesome-agentic-ai-zh

WenyuChiou/awesome-agentic-ai-zh · ★ 7.0k ▲281/7d · Python

awesome-agentic-ai-zh is a Traditional Chinese-first learning roadmap for AI agents, with complete English and Simplified Chinese editions maintained alongside it. It organizes concepts, projects, and small exercises into separate paths for using agent CLIs and building agent systems.

Verdict Our run installed 35 packages in 14 seconds, built the documentation in 4 seconds, and found 0 known vulnerabilities, making the curriculum easy to inspect locally. Choose it if you want a maintained, trilingual route through CLI agents, Claude Code, MCP, and multi-agent systems.

kubernetes-mcp-server

containers/kubernetes-mcp-server · ★ 2.1k ▲22/7d · Go

kubernetes-mcp-server lets MCP clients inspect and operate Kubernetes and OpenShift through a native Go server. It talks directly to the Kubernetes API and exposes tools for resources, pods, logs, metrics, Helm, multiple clusters, and optional ecosystem integrations.

Verdict Our kubernetes-mcp-server run installed 393 Go packages, built successfully, and passed all 36 tests in 102 seconds, the strongest clean build result in this batch. Use it for supervised cluster diagnosis with a dedicated read-only ServiceAccount, a narrow toolset, and sensitive kinds denied.

codex

openai/codex · ★ 124k ▲1.9k/7d · Rust

Codex CLI is OpenAI's local coding agent for working on repositories from a terminal. It can inspect files, edit code, run commands under sandbox and approval rules, and connect the work to a ChatGPT account or OpenAI API key.

Verdict Our Codex checkout installed 1 pnpm package in 21 seconds but exposed no build or test target, so that run validates only a thin Node workspace path and not the 1.5-million-line Rust project. The released CLI is worth using for developers who want an agent inside a real terminal with visible edits, commands, and approval boundaries.

Front-End-Checklist

thedaviddias/Front-End-Checklist · ★ 74k ▲63/7d · MDX

Front-End Checklist is a catalog of 385 English rules for reviewing HTML, CSS, JavaScript, accessibility, performance, SEO, security, images, testing, privacy, and internationalization. Humans can browse it, while AI agents can query the same material through a hosted MCP server or installable skills.

Verdict Our Front-End Checklist run installed 1,362 packages and used 1,421 MB, then built in 105 seconds and passed all 20 Jest tests in 29 seconds. Use the README or hosted rule browser as a strong review prompt, and add the MCP only when agent access to 385 rules justifies an external service connection.

docling

docling-project/docling · ★ 66k ▲243/7d · Python

Docling turns PDFs, office files, images, audio, video, email, and other formats into structured Markdown, HTML, JSON, or its own document model. It handles the messy first step in search and AI systems: recovering text, tables, layout, reading order, and media while allowing local processing.

Verdict Docling is the first project we would trial when one local pipeline must understand many document types and preserve more than plain text. Its format range, document model, exports, integrations, service, and MCP route save substantial assembly work.

litellm

BerriAI/litellm · ★ 59k ▲515/7d · Python

LiteLLM is an open-source Python SDK and deployable gateway that presents many model providers through a mostly OpenAI-shaped API. It gives applications one place for routing, fallbacks, virtual keys, budgets, usage records, guardrails, and provider credentials instead of baking each vendor's client into every service.

Verdict LiteLLM is the practical default when a team needs one API in front of several model vendors. The provider coverage and proxy controls solve real duplication, but the gateway becomes a privileged billing and credential boundary that deserves careful operation.

chrome-devtools-mcp

ChromeDevTools/chrome-devtools-mcp · ★ 52k ▲682/7d · TypeScript

Chrome DevTools MCP is a Model Context Protocol server and standalone CLI that lets coding agents control and inspect a live Chrome browser. It exposes page automation, screenshots, console and network data, performance traces, PWA checks, and memory analysis through Chrome DevTools and Puppeteer.

Verdict Our Chrome DevTools MCP build finished in 43 seconds, but its test suite failed after 108 seconds in a Puppeteer browser-connection path even though npm audit found 0 vulnerabilities. Use it when an agent needs real console, network, trace, PWA, or heap evidence from Chrome.

QwenPaw

agentscope-ai/QwenPaw · ★ 35k ▲263/7d · Python

QwenPaw is a self-hosted personal AI assistant that keeps one agent, its memory, tools, and conversations available through a browser, terminal, desktop app, and chat services. It solves the glue-work problem of connecting models, scheduled jobs, MCP tools, and messaging channels without building that control plane yourself.

Verdict Our QwenPaw install took 108 seconds, placed 277 packages and 999 MB on disk, then ran tests for 900 seconds without finishing, so this is a serious self-hosted agent platform rather than a light chat wrapper. Use it when persistent memory, messaging channels, MCP, scheduling, and local models belong in one system and you can operate the permissions around them.

agentscope

agentscope-ai/agentscope · ★ 32k ▲518/7d · Python

AgentScope is a Python framework for building tool-using AI agents, then running them as multi-user services with storage, scheduling, retrieval, and chat interfaces. It solves the jump from a small agent loop to an application that needs sessions, permissions, workspaces, and several model providers.

Verdict Our AgentScope run passed 1,977 tests after a 269-second install that consumed 1,039 MB, so it is well tested and expensive to install. Use it when your Python agent needs to grow into a multi-user service with permissions, persistence, MCP, and isolated workspaces.

serena

oraios/serena · ★ 29k ▲366/7d · Python

Serena is an MCP server that gives coding agents symbol-aware search, editing, refactoring, and project memory through language servers or a paid JetBrains plugin. It tackles the cases where plain text search and line-based edits are too crude for following references or changing code across a large repository.

Verdict Our Serena run built successfully, but the 840-second test step ended with 36 failures and 164 setup or collection errors, including an explicit missing-Ruby error. It is still worth trying when an MCP coding agent regularly loses track of symbols across a large repository, because 1,253 tests passed and the free backend covers more than 40 listed languages.

Chat2DB

OtterMind/Chat2DB · ★ 28k ▲28/7d · Java

Chat2DB Community is a local-first database client with an English README and translations in Chinese, Japanese, Spanish, and Korean. It combines a SQL editor, database object management, import and export, charts, and a bring-your-own-model assistant that can generate, explain, and optimize SQL.

Verdict Our Chat2DB frontend install pulled 3,168 packages and used 1,635 MB, then built in 98 seconds with no test target available. It is worth trying as a single-user database workbench when broad driver coverage and optional AI assistance matter, preferably through the desktop package.

python-sdk

modelcontextprotocol/python-sdk · ★ 24k ▲51/7d · Python

The MCP Python SDK is the official Python library for building Model Context Protocol servers and clients. It turns typed Python functions into tools, resources, and prompts, then handles stdio, Streamable HTTP, or SSE communication with compatible AI applications.

Verdict Our MCP Python SDK run installed 61 packages and built in 38 seconds combined, but its tests never collected because `logfire` was absent. Use it when Python and official protocol coverage matter, especially for a server or client that must track the current MCP specification.

go-micro

micro/go-micro · ★ 23k ▲8/7d · Go

Go Micro lets Go developers build ordinary network services, AI agents, and repeatable workflows in the same runtime. It can expose service methods as MCP tools, connect agents over A2A, and keep model-driven steps beside code paths whose order is fixed.

Verdict Our Go Micro run built in 102 seconds and passed all 162 tests in 97 seconds, so v6 is worth a serious trial for Go teams that want agents and services in one process model. Use the deterministic flow API for business steps and place approval checks around model-selected tools.

typescript-sdk

modelcontextprotocol/typescript-sdk · ★ 13k ▲46/7d · TypeScript

The official MCP TypeScript SDK supplies the client and server pieces for connecting AI applications to tools, resources, and prompts. It handles the protocol plumbing across Node.js, Bun, and Deno, including stdio, Streamable HTTP, OAuth helpers, and adapters for several web frameworks.

Verdict Our TypeScript SDK install took 33 seconds and 665 MB, but the checkout offered no root build or test target, so adoption starts with good protocol coverage and an incomplete top-level verification story. Use it for serious TypeScript MCP work because it is the official implementation and tracks the 2026-07-28 specification closely.

browser-tools-mcp

AgentDeskAI/browser-tools-mcp · ★ 7.3k ▲4/7d · TypeScript

BrowserTools MCP lets an AI coding client inspect the Chrome tab a developer is already using. A local MCP server and unpacked extension expose console output, network traffic, selected elements, screenshots, browser storage, and Lighthouse audits without moving the session into an automated browser profile.

Verdict Our browser-tools-mcp run built in 8 seconds and passed 355 of 356 tests, with the one failure in degraded attachment handling and 0 audit findings. It is a strong choice for debugging the authenticated Chrome tab in front of you, especially when console and network evidence matter more than browser automation.

genkit

genkit-ai/genkit · ★ 6.4k ▲28/7d · TypeScript

Genkit is an open-source framework for building server-side AI features in JavaScript and TypeScript, Go, Python, or Dart. It gives application developers one interface for model calls, structured output, tools, retrieval, workflows, local debugging, and production traces instead of making them assemble each part separately.

Verdict Our Genkit build succeeded, but it took 407 seconds after an 8-second install, so source contributors should budget for a slow loop even though dependency setup was quick. Use Genkit when an application team wants typed AI workflows, provider plugins, local traces, and a documented route into Firebase or Cloud Run.

apfel

Arthur-Ficial/apfel · ★ 6.4k ▲26/7d · Swift

apfel is a command-line tool for macOS that unlocks the free, on-device AI model built into Apple Silicon Macs. It lets you use this powerful, private AI directly in your terminal for scripting and automation, or run it as a local server that mimics the OpenAI API for your development projects.

Verdict apfel is a must-have utility for any developer or power user on a modern Mac. It brilliantly bridges the gap between Apple's powerful, private on-device AI and the command-line workflows where it can be most useful.

gemini-cli

google-gemini/gemini-cli · ★ 107k ▲133/7d · TypeScript

Gemini CLI is Google's open-source terminal agent for reading code, editing files, running shell commands, searching the web, and calling Gemini models. It can work interactively or in scripts, and it supports MCP servers, project instructions, checkpoints, IDE connections, and a GitHub Action.

Verdict Our Gemini CLI run installed 1,325 packages and built successfully, but 10 of 7,000 tests failed before the suite hit 900 seconds and npm audit reported 71 vulnerabilities. It is worth trying for developers committed to Gemini who need a capable terminal and MCP client, especially through the published package.

awesome-mcp-servers

punkpeye/awesome-mcp-servers · ★ 95k ▲401/7d

This is not a single piece of software, but a massive, curated directory of open-source servers built on the Model Context Protocol (MCP). MCP is a standard that lets AI models and agents safely connect to external tools like databases, APIs, and local files. This 'awesome list' solves the critical discovery problem for developers, providing a central catalog for finding pre-built capabilities to give their AI agents.

Verdict Awesome MCP Servers is the definitive, indispensable catalog for the agentic AI revolution. While its massive popularity has led to clear maintenance challenges, evidenced by the thousands of open issues, its comprehensive scope and excellent organization make it the essential first stop for any developer looking to empower their AI with real-world capabilities.

netdata

netdata/netdata · ★ 81k ▲68/7d · Go

Netdata is an English-language infrastructure monitoring system that collects per-second metrics, stores and analyzes them on the monitored machines, and presents ready-made dashboards and alerts. Its agent covers hosts, containers, services, logs, anomaly detection, and an MCP interface, while optional Netdata Cloud adds shared views and access control.

Verdict Our measured Netdata subproject built in 17 seconds, but pytest collected 0 tests and pip-audit found 1 known vulnerability, so that run cannot support a clean bill of health for the monitoring agent. Netdata is worth a trial when quick per-second host visibility and ready-made dashboards matter more than adopting a query-centered metrics stack.

context7

upstash/context7 · ★ 62k ▲273/7d · TypeScript

Context7 retrieves current library documentation and code examples for coding agents through a CLI, an MCP server, or an SDK. It is meant to stop an agent from relying on old API details when answering a version-specific programming question.

Verdict Our Context7 install and build finished in 38 seconds combined, but the 10-second test command still exited 1 in the SDK workspace. It is worth using when stale library APIs repeatedly waste agent turns and a hosted documentation index is acceptable.

mempalace

MemPalace/mempalace · ★ 59k ▲118/7d · Python

MemPalace is a local-first memory store for AI assistants that keeps original conversation text and retrieves it with semantic and lexical search. It can mine project files and chat transcripts, organize them into scoped collections, expose memory through MCP, and save sessions from Claude Code, Codex CLI, and Cursor hooks.

Verdict MemPalace is a thoughtful fit for developers who want the original record kept locally and searchable across coding-agent sessions. Its isolated install and default embedded backend are approachable, but the 15-minute test timeout, one reported dependency vulnerability, and multi-writer edge cases argue against casual team-wide deployment.

CowAgent

zhayujie/CowAgent · ★ 47k ▲149/7d · Python

CowAgent is a self-hosted assistant that can plan work, use files and a terminal, browse the web, remember past conversations, and reply through chat services. It puts those jobs behind one web console instead of making you assemble a separate bot, memory store, scheduler, and tool runner. The default README is in English, with Chinese, Traditional Chinese, and Japanese translations.

Verdict Our CowAgent run installed 98 packages in 48 seconds, but 77 tests failed and pip-audit found 35 known vulnerabilities, so this is a capable personal-agent stack that needs a guarded trial. Use it when one assistant genuinely must span memory, tools, skills, MCP, and several chat services.

agentic-awesome-skills

sickn33/agentic-awesome-skills · ★ 46k ▲257/7d · Python

Agentic Awesome Skills is a large catalog of reusable instructions for coding agents, plus local tools for finding, checking, and recording a chosen set. It helps Codex, Claude Code, Cursor, and other supported clients reuse task-specific playbooks without copying prompts by hand.

Verdict Our sandbox installed 11 packages in 7 seconds but could not build because the Python validator lacked `yaml`, so Agentic Awesome Skills is easier to consume than to develop from a bare Node container. Use a selected skill or focused plugin if you need breadth across coding agents and will review what the agent chooses.

agno

agno-agi/agno · ★ 42k ▲74/7d · Python

Agno is a Python framework and runtime for building agents, serving them through an API, and operating them from a web interface. It bundles storage, traces, permissions, integrations, scheduling, and several agent protocols so a team can own more of its agent platform.

Verdict Our Agno quickstart installed 135 packages into 482 MB and pip-audit found 3 known vulnerabilities, so adopting it means accepting a real platform dependency, not adding a tiny agent helper. It is a strong candidate for Python teams that need agents, APIs, storage, permissions, and operations in one system.

chatbox

chatboxai/chatbox · ★ 42k ▲71/7d · TypeScript

Chatbox Community Edition is a desktop and mobile client for talking to OpenAI, Anthropic, Google, local Ollama models, and other providers from one interface. It stores conversations locally, renders Markdown, code, math, and images, and adds prompt management, model switching, streaming, and MCP-related tooling around provider APIs.

Verdict Our Chatbox build finished in 108 seconds and 2,692 of 2,780 tests passed, but 2 context-related tests still failed, so the pinned source was close to green rather than green. Chatbox is a good choice for an individual who wants one local desktop history across hosted and Ollama models.

claude-howto

luongnv89/claude-howto · ★ 42k ▲98/7d · Python

Claude How To is an English-language visual curriculum for learning Claude Code, with Vietnamese, Chinese, Ukrainian, and Japanese translations. It organizes commands, memory, skills, subagents, MCP, hooks, plugins, checkpoints, and CLI use into a guided path with templates you can copy into a real project.

Verdict Our run installed 35 packages in 13 seconds, built in 5 seconds, and exposed no test target, so Claude How To is easy to inspect but cannot validate itself through the path our harness found. Use it as a guided course and a box of editable starting points, especially if the official reference feels fragmented.

AstrBot

AstrBotDevs/AstrBot · ★ 41k ▲314/7d · Python

AstrBot is an open-source Python platform for putting an AI assistant inside chat systems such as Telegram, Slack, Discord, QQ, Feishu, and DingTalk. Its main README is in English, with Chinese, Japanese, French, Spanish, and Russian versions linked, and its documentation and release notes include substantial Chinese-language material.

Verdict Our AstrBot install and build finished in 82 seconds total, and pip-audit found 0 known vulnerabilities, but tests never started because `pytest_asyncio` was missing. AstrBot is a strong candidate for a multilingual, self-hosted assistant spanning several chat networks, especially the Chinese platforms its official adapters cover.

DeepTutor

HKUDS/DeepTutor · ★ 40k ▲611/7d · Python

DeepTutor is a self-hosted learning workspace that combines chat, document retrieval, guided study, quizzes, writing, books, and persistent AI partners. The main README is English and links translated documentation, while users bring their own local or hosted model providers.

Verdict Our DeepTutor run installed 200 packages and built successfully, but pytest collected 0 tests because configuration and a test dependency were missing, while pip-audit found 4 known vulnerabilities. That makes it a strong evaluation candidate for technical learners, not a release you should approve from the green build alone.

claude-code-router

musistudio/claude-code-router · ★ 37k ▲95/7d · TypeScript

Claude Code Router is a local gateway that puts several coding agents behind one provider, model, credential, routing, and logging layer. Its English README is the primary page and a Chinese translation is available; the app supports Claude Code, Codex, OpenCode, Kimi CLI, and other compatible clients while translating requests across several model API formats.

Verdict Our Claude Code Router install pulled 1,892 packages and 493 MB, then the release build stopped at 32 seconds for a missing `GH_TOKEN`, so this is a substantial control plane rather than a tiny proxy. Use it when several agents and providers already create enough configuration churn to justify one local routing and observability layer.

claude-plugins-official

anthropics/claude-plugins-official · ★ 36k ▲272/7d · Python

Claude Plugins Official is Anthropic's managed marketplace directory for Claude Code plugins. It contains Anthropic-maintained plugins and external entries from partners or the community, with optional commands, agents, skills, language servers, hooks, and MCP server connections.

Verdict Our Discord-plugin install took 46 seconds for 130 packages, yet the directory offered no build or test target for that measured path. Use Claude Plugins Official as a discovery index and convenient installer, not as a blanket trust decision.

DeepSeek-Reasonix

esengine/DeepSeek-Reasonix · ★ 36k ▲106/7d · Go

Reasonix is a Go-based coding agent that runs in a terminal, desktop app, browser, or editor through ACP. It gives long coding sessions a shared engine with permissions, workspace boundaries, checkpoints, model choice, and plugins, so developers can leave work running without surrendering every file and command to an opaque process.

Verdict Our 749-second test run ended with 25 of 26 Go packages passing, so Reasonix is buildable but not clean enough at commit b9cf32f for an unquestioned rollout. Use it if its shared CLI, desktop, web, and ACP engine solves a real workflow problem and you are prepared to pin a release.

Vibe-Trading

HKUDS/Vibe-Trading · ★ 34k ▲446/7d · Python

Vibe-Trading is a Python research workspace that connects an AI agent to market-data loaders, backtests, portfolio analysis, reports, broker journals, finance tools, and optional broker connectors. It helps turn a plain-language trading question into inspectable code and artifacts, while also exposing research tools through a CLI, web app, REST API, messaging channels, and MCP.

Verdict Our Vibe-Trading run passed 11,133 of 11,134 tests, with the only failure caused by a missing passwd entry for container UID 1000. That is strong evidence for its research code, and none for expected returns or safe live execution.

AIHawk

feder-cr/AIHawk · ★ 32k ▲1.3k/7d · Python

AIHawk is an open-source AI browser agent: you tell it in plain language to browse, click, type and read real web pages, and it drives an actual browser to do it. The problem it solves is slow, brittle, manual web workflows, like checking fares across calendar widgets, or filling forms that lack APIs.

Verdict AIHawk installed in 39 seconds, passed all 138 tests in 22 seconds, and exposed zero known vulnerabilities in our sandbox; if you can accept the OpenRouter key and the fact that a browser agent's judgment is only as good as the model you point it at, it is a legitimate shortcut for real web automation..

blender-mcp

MCPBlender/blender-mcp · ★ 29k ▲914/7d · Python

Blender MCP connects an MCP-capable AI client to a Blender add-on so the model can inspect a scene, edit objects and materials, run Python, and bring in external assets. It solves the awkward handoff between a text agent and Blender by translating MCP tool calls into commands over a local socket.

Verdict Our Blender MCP run installed 60 packages and passed all 20 tests in 35 seconds across install, build, and test steps. Use it for supervised scene prototyping where a technical artist can inspect each change and revert the Blender file.

blender-mcp

ahujasid/blender-mcp · ★ 29k ▲914/7d · Python

Blender MCP connects an MCP-capable AI client to a Blender add-on so the model can inspect a scene, edit objects and materials, run Python, and bring in external assets. It solves the awkward handoff between a text agent and Blender by translating MCP tool calls into commands over a local socket.

Verdict Our Blender MCP run installed 60 packages and passed all 20 tests in 69 seconds across install, build, and test steps. Use it for supervised scene prototyping where a technical artist can inspect each change and revert the Blender file.

mlflow

mlflow/mlflow · ★ 28k ▲100/7d · Python

MLflow is a server and SDK for recording machine-learning experiments, registering models, tracing agent calls, evaluating outputs, and controlling access to model providers. It gives teams one place to inspect runs and production AI behavior instead of assembling separate tracking, registry, and observability services.

Verdict Our MLflow run installed 116 packages and used 601 MB, then 273 tests passed before 200 collection or setup errors stopped the suite. Use it when experiment tracking, model governance, and agent observability genuinely belong on one operated platform.

FunASR

modelscope/FunASR · ★ 20k ▲119/7d · Python

FunASR is a Python speech-recognition toolkit for offline transcription, streaming recognition, voice activity detection, punctuation, speaker diarization, emotion tags, training, and model deployment. It offers several checkpoint families instead of one fixed model, plus an OpenAI-compatible server and an MCP server for agent clients. The default README is English, with Simplified Chinese, Japanese, and Korean versions linked at the top.

Verdict Our FunASR run installed 94 packages and built in 13 seconds, but pytest executed 0 tests because collection ended with a `SystemExit: 2` error. Trial it when Chinese or multilingual speech, streaming, diarization, or edge deployment requires more choice than one Whisper model provides.

SkillSpector

NVIDIA/SkillSpector · ★ 17k ▲652/7d · Python

SkillSpector checks AI-agent skill files for prompt injection, data theft, dangerous code, excessive permissions, vulnerable dependencies, and other install risks. It reads the skill without running it, then combines static checks with optional language-model analysis to produce a score and recommendation.

Verdict Our run installed 103 npm packages in 35 seconds and used 232 MB, but the checkout exposed no build or test target, so it did not verify the Python scanner. Use SkillSpector as pre-install triage alongside coverage checks and human review, especially for skills with scripts or broad permissions.

claude-seo

AgriciDaniel/claude-seo · ★ 17k ▲323/7d · Python

Claude SEO is a Claude Code plugin that turns website auditing into a set of local commands and specialist agent workflows. It checks technical SEO, content, schema, search data, local visibility, and AI citation signals, then writes reports and prioritized actions instead of locking the findings in a vendor dashboard.

Verdict Our Claude SEO run installed 92 packages and built cleanly, but 2 tests failed when a dry-run hit an HTTP 403 rate limit. Use it as an open, inspectable audit assistant if Claude Code is already part of the job and a human will verify the recommendations.

xiaohongshu-mcp

xpzouying/xiaohongshu-mcp · ★ 16k ▲110/7d · Go

xiaohongshu-mcp is a Chinese-first MCP server that lets an AI client operate Xiaohongshu through an automated browser. A maintained English README exists. Its tools can log in, search and read posts, inspect profiles, publish images or video, and perform account actions such as commenting, liking, and saving.

Verdict Our xiaohongshu-mcp run installed 67 packages in 14 seconds and passed all 8 test packages, but that clean result cannot protect browser selectors from a Xiaohongshu page change. Use it for supervised local workflows where a person approves every public action and can repair login or search when the site moves.

Graft

NanoNets/Graft · ★ 8.1k ▲1.4k/7d · TypeScript

Graft builds a local graph of a codebase so coding agents can find systems, symbols, callers, and change impact without rereading the repository from scratch. Its structural layer uses tree-sitter, while an optional model-backed pass adds summaries and concept nodes for Claude Code, Codex, Cursor, Gemini, and other agent hosts.

Verdict Our Graft run built in 7 seconds and passed all 836 tests, but npm audit found 1 high-severity vulnerability, so the code earns a serious trial with a dependency review attached. Use it when coding agents repeatedly lose time mapping the same large repository and your team can inspect the generated context and configuration changes.

firecrawl-mcp-server

firecrawl/firecrawl-mcp-server · ★ 7.5k ▲35/7d · JavaScript

Firecrawl MCP Server lets an MCP-compatible agent search the web, scrape pages, crawl sites, parse files, and interact with pages through Firecrawl. It solves the adapter work between an AI client and Firecrawl's hosted or self-hosted API, with tool schemas, transports, authentication, retries, and rate handling already wired.

Verdict Our Firecrawl MCP Server run installed 341 packages, built in 5 seconds, and passed all 74 tests in 24 seconds, making the adapter easy to trial. Use it when Firecrawl is already your web-data service or when its hosted endpoint is an acceptable dependency.

gptme

gptme/gptme · ★ 4.4k ▲4/7d · Python

gptme is a local terminal agent that can edit files, run shell and Python commands, browse the web, inspect images, call MCP tools, and keep conversations across sessions. It supports several hosted model providers plus local models, and it can run interactively, from an editor through ACP, or as a scheduled headless agent.

Verdict Our gptme run passed 9,557 tests but still ended with 27 failures and 75 errors after 632 seconds, so the project is substantial and the current checkout needs environment-specific verification. It is a strong fit for an experienced terminal user who wants provider choice, local state, MCP, and deep customization.

Unity-MCP

IvanMurzak/Unity-MCP · ★ 4.3k ▲49/7d · C#

Unity-MCP connects an MCP-capable coding agent to the Unity Editor and, with custom tools, to a compiled game. It lets the agent inspect scenes, edit scripts and assets, run tests, read profiler data, and call project-specific C# methods instead of merely suggesting code in chat.

Verdict Our Unity-MCP CLI run installed 64 packages, built in 12 seconds, and passed its tests in 33 seconds, making it an easy technical trial for a Unity team. Use it when you want an agent acting inside the editor and you will inspect every mutation in version control.

sie

superlinked/sie · ★ 3.3k ▲36/7d · Python

Superlinked Inference Engine, or SIE, is a self-hosted server that puts embeddings, reranking, document conversion, structured extraction, safety checks, and open language models behind one API. It is meant to replace a pile of separate model services with a catalog that loads models as jobs need them.

Verdict Our SIE run installed 704 packages and used 1,597 MB, then passed its build and tests in 31 seconds combined, so the codebase is workable but hardly small. Use it when several private model services have become one platform problem and you already have people who can operate the cluster.

harbor

av/harbor · ★ 3.2k ▲6/7d · Python

Harbor is a CLI and desktop companion for assembling local AI services with Docker Compose. It connects model backends, chat interfaces, search, speech, image generation, MCP tools, and coding agents so you can try a stack without wiring every container by hand.

Verdict Our Harbor checkout installed 97 packages in 19 seconds and used 121 MB, but it offered no build or test target, so the easy setup does not provide a full verification result. Use Harbor to explore and compare a local AI stack, especially when hand-wiring Compose services is slowing experimentation.

webbrain

webbrain-one/webbrain · ★ 1.1k ▲62/7d · JavaScript

WebBrain is an open-source browser extension that lets an AI model read pages, answer questions, and carry out browser tasks in the tabs you already use. It can use its managed model, a cloud API, or a local OpenAI-compatible server, and it also lets coding agents delegate browser work through MCP.

Verdict Our WebBrain checkout installed 4 packages in 11 seconds and passed its available tests in 74 seconds, with 0 npm audit findings. Use it when an agent must work in the signed-in browser session you already have, especially through a local model or MCP client.

firecrawl

firecrawl/firecrawl · ★ 181k ▲2.5k/7d · TypeScript

Firecrawl is a web data API that can search, scrape, map, crawl, and interact with pages, then return markdown, structured JSON, screenshots, or page metadata. It gives agents and applications one interface for web work that may require JavaScript rendering or browser actions before extraction.

Verdict Our Firecrawl build failed in 9 seconds on an unresolved Rust-backed module, and 183 of 2,166 tests failed, so self-hosters should treat adoption as a platform project. The hosted API is the sensible first trial for teams that want search, crawling, extraction, and browser actions behind one interface.

firecrawl

mendableai/firecrawl · ★ 181k ▲2.8k/7d · TypeScript

Firecrawl is an API and self-hostable service that searches the web, turns pages into markdown or structured data, crawls sites, and controls browser sessions. It gives agents and applications one interface for pages that may need JavaScript rendering, link discovery, document parsing, or interaction before extraction.

Verdict Firecrawl is a convincing hosted choice when a product needs search, scraping, crawling, and browser interaction behind one API. The open code is useful too, but our unresolved Rust package, 183 failed tests, and multi-service development path make self-hosting a platform job rather than a cheap escape from API fees.

skills

anthropics/skills · ★ 177k ▲1.1k/7d · Python

Anthropic Skills is a collection of instruction folders, scripts, examples, and a specification that teach Claude how to perform repeatable tasks. It includes document creation skills, technical examples, a starter template, and plugin packages for Claude Code.

Verdict Anthropic Skills is a useful reference shelf, especially for seeing how a short instruction file can expand into scripts and supporting material. It is not a vetted package set to install wholesale.

ponytail

DietrichGebert/ponytail · ★ 139k ▲6.0k/7d · JavaScript

Ponytail is an English coding-agent skill that pushes an agent to solve a task with less new code: reuse the codebase, standard library, browser, platform, and installed dependencies before creating another abstraction. Spanish and Korean README translations exist. Plugins and rule-file adapters bring that guidance to Claude Code, Codex, Copilot CLI, OpenCode, Gemini, Hermes, and many other agent hosts.

Verdict Ponytail is a sharp corrective for agents that turn a date input into a component project. Its decision ladder is easy to understand, the adapters are unusually broad, and the repository stays small.

graphify

Graphify-Labs/graphify · ★ 118k ▲2.0k/7d · Python

Graphify turns a repository and its supporting documents into a local knowledge graph that an engineer or coding assistant can query. It parses code into named nodes and relationships, then produces an interactive HTML graph, a Markdown report, and JSON for path, explanation, and natural-language queries.

Verdict Our Graphify run passed 4,541 tests, but 24 Terraform cases failed because `tree_sitter_hcl` was unavailable, so the base tool looks well exercised while optional-language packaging still needs verification. Use it for faster orientation and scoped assistant context in a large repository, provided every consequential edge is checked against source.

caveman

JuliusBrussee/caveman · ★ 106k ▲1.4k/7d · Go

Caveman is a set of local tools that tries to cut the text coding agents read and write. Its skill asks agents for shorter answers, while its proxy compresses request context and keeps recoverable copies of the original data on your machine.

Verdict Our Caveman install pulled 391 packages and used 915 MB, then its test command stopped after 4 seconds on a Node option the runtime rejected. Try the skill first if shorter agent answers are the goal; adopt the proxy only after checking output quality, credential routing, telemetry, and the BSL terms.

servers

modelcontextprotocol/servers · ★ 90k ▲200/7d · TypeScript

This repository contains seven small reference servers that demonstrate how Model Context Protocol clients can expose files, Git repositories, web fetching, memory, time, prompts, resources, and tools to a model. It is an educational code collection maintained by the MCP steering group, not the general MCP server directory or a production integration bundle.

Verdict Our build finished in 18 seconds, but no test target existed and npm audit found 5 known vulnerabilities, including 4 high-severity advisories. Use this repository to learn MCP shapes or test a client, exactly as its maintainers intend.

NextChat

ChatGPTNextWeb/NextChat · ★ 89k ▲52/7d · TypeScript

NextChat is a self-hosted chat interface for OpenAI, Anthropic, Google, DeepSeek, Azure, and compatible model APIs. It gives individuals or small teams one browser and desktop client for conversations, prompt templates, artifacts, plugins, and optional MCP tools while keeping chat data in the browser by default.

Verdict NextChat is a good fit for a personal or small-team model console when browser-local storage, provider choice, and quick deployment matter. The open-source edition is much weaker as an organizational service because its documented access control is a shared password, and current proxy security reports need an operator's attention.

worldmonitor

koala73/worldmonitor · ★ 86k ▲626/7d · TypeScript

World Monitor is a global intelligence dashboard that combines news, maps, markets, geopolitical signals, infrastructure data, and AI-generated briefs. It gives analysts and curious readers one situational view instead of making them watch many feeds and specialist sites separately.

Verdict Our World Monitor install consumed 2,346 MB, npm audit found 38 known vulnerabilities, and the repository exposed no test target, even though its 112-second production build passed. Use it as an exploratory intelligence desk when breadth, maps, and agent access outweigh a heavy operating surface.

Agent-Reach

Panniantong/Agent-Reach · ★ 82k ▲3.2k/7d · Python

Agent Reach's primary README is Chinese, and a substantial English translation is available in the repository. It gives command-running AI agents a maintained routing layer for reading and searching websites, social networks, video platforms, GitHub, and feeds through selected upstream tools.

Verdict Our Agent Reach run installed 45 packages, used 66 MB, and completed its test command successfully, so the core is cheap to trial. Use it when an agent needs several web and social sources and you accept that cookies, browser sessions, and upstream breakage remain your operational responsibility.

headroom

headroomlabs-ai/headroom · ★ 72k ▲1.5k/7d · Python

Headroom compresses tool output, logs, retrieved text, files, and conversation history before an AI model reads them. It can sit inside Python or TypeScript code, run as a local proxy, or expose compression and retrieval through MCP, with original content kept locally for later retrieval.

Verdict Our Headroom install pulled 218 packages and occupied 6,039 MB, while its test suite was still at 30% when our 900-second cap stopped it. Trial it if long agent sessions are expensive enough to justify a local proxy and you can compare compressed answers with the originals.

strix

usestrix/strix · ★ 63k ▲1.2k/7d · Python

Strix is an open-source security testing agent that probes source code, web apps, and APIs, then records vulnerabilities with reproduction steps. It is meant to give developers a practical pentest loop without hiring a human tester for every code change, though it does not replace an independent assessment for high-risk systems.

Verdict Our Strix run passed 390 tests but failed 200 before pytest stopped, and pip-audit reported 3 known vulnerabilities. It is worth a controlled trial for security teams that understand pentesting and want an agent to widen routine coverage.

goose

block/goose · ★ 54k ▲283/7d · Rust

Goose is a local AI agent with a desktop app, command-line interface, and embeddable API for editing code, running tools, research, and repeatable workflows. The original `block/goose` address now resolves to `aaif-goose/goose`, where the project supports hosted and local model providers plus extensions through MCP.

Verdict Our Goose build and test commands each exceeded 900 seconds, so the packaged app is easier to recommend than source contribution on a 3-CPU machine. Use Goose when provider choice and MCP extensions are worth a powerful local agent, but approve its commands and inspect every shared recipe.

nanobot

HKUDS/nanobot · ★ 48k ▲295/7d · Python

Nanobot is a self-hosted personal AI agent that runs through a browser, terminal, API, or chat app. It gives one model access to files, shell commands, web tools, memory, MCP servers, scheduled jobs, and helper agents, so a developer can operate an assistant on their own machine or server.

Verdict Our sandbox installed Nanobot in 45 seconds and built it in 9, but 195 tests failed and pip-audit found 42 known vulnerabilities, so we would not deploy commit 3852956 without investigation. It is still a strong trial for experienced self-hosters who want one agent across a browser, terminal, API, and chat services.

codebase-memory-mcp

DeusData/codebase-memory-mcp · ★ 43k ▲718/7d · C

Codebase Memory MCP indexes source repositories into a local knowledge graph that coding agents can query through Model Context Protocol tools. It gives an agent structural search, call paths, change-impact maps, architecture summaries, code snippets, and cross-repository links without sending source code to a hosted indexing service.

Verdict Our graph UI build passed in 15 seconds and its tests passed in 10 seconds after a 352-package install, but we did not execute the native indexer those screens control. Codebase Memory MCP is worth piloting for local, multi-repository agent navigation when file-by-file exploration wastes context.

CodeWhale

Hmbown/CodeWhale · ★ 41k ▲51/7d · Rust

CodeWhale is a Rust coding agent that runs in a terminal, reads a repository, edits files, and executes commands with configurable approval rules. It supports hosted providers and local models, while keeping sessions, goals, agent roles, MCP connections, and recovery controls in one TUI.

Verdict Our CodeWhale install took 13 seconds and found 0 known npm vulnerabilities, but the package offered no build or test target, so the easy install is stronger evidence than the local verification story. Choose it if provider choice, resumable goals, MCP, and explicit access modes are central requirements.

playwright-mcp

microsoft/playwright-mcp · ★ 37k ▲245/7d · TypeScript

Playwright MCP is a Model Context Protocol server that lets an AI agent browse and operate websites through Playwright. It sends structured accessibility snapshots to the model, giving the agent named elements and stable references for navigation, forms, screenshots, storage, network work, and test assertions.

Verdict Our Playwright MCP build took 6 seconds, but its tests stopped in another 6 seconds because the process used Node 18.20.8 and Playwright required Node 20 or newer. Use it when an MCP agent needs persistent, inspectable browser state and you accept the token and credential costs.

PageIndex

VectifyAI/PageIndex · ★ 36k ▲51/7d · Python

PageIndex builds a hierarchical tree from a document and lets a language model navigate that structure to answer questions, instead of splitting text into fixed chunks and searching a vector database. Its Python SDK can index and chat locally with your own model connection, or use a hosted service for OCR, image understanding, line-level citations, and corpus-scale features.

Verdict Our PageIndex run passed 345 tests with 0 failures in 181 seconds, while 50 tests were skipped and pip-audit found 1 vulnerability. Try it when long, structured PDFs defeat chunk-and-vector retrieval and you can pay for or host the model calls that tree search requires.

VoiceStudio

debpalash/VoiceStudio · ★ 31k ▲9.1k/7d · Python

VoiceStudio is a local desktop and web workspace for cloning voices, generating speech, transcribing recordings, dubbing video, dictating text, and producing long-form audio. It brings several speech engines into one interface, so creators can keep core processing and project files on their own machines instead of paying for a hosted voice service.

Verdict VoiceStudio is a convincing choice for a technically comfortable creator who wants several private speech workflows in one local app. The successful install and build are encouraging, though the 849 MB dependency footprint, model downloads, platform rules, and current beta reports make this a machine to tend rather than an appliance to forget.

gbrain

garrytan/gbrain · ★ 30k ▲196/7d · TypeScript

GBrain is a self-hosted memory and knowledge layer for AI agents, built around Markdown files, a graph, search, synthesis, and background maintenance. It connects to Codex, Claude Code, OpenClaw, Hermes, and other MCP clients so an agent can recall information across sessions instead of relying on one chat history.

Verdict Our GBrain run built in 7 seconds and Bun counted 7,152 passing tests, but the full command still exited 1 after 869 seconds, so this commit did not clear its test gate. GBrain is compelling for a technical user who wants Git-backed agent memory and can own privacy, providers, and upgrades.

xiaozhi-esp32

78/xiaozhi-esp32 · ★ 30k ▲222/7d · C++

XiaoZhi ESP32 is firmware for building a voice assistant on ESP32 boards, with audio, displays, cameras, networking, and device control handled on the microcontroller. It connects the device to speech and language-model services, while MCP lets the assistant operate hardware such as speakers, LEDs, servos, and GPIO pins.

Verdict Our XiaoZhi ESP32 setup failed in 1 second because plain CMake could not load the ESP-IDF project file or recognize an ESP-IDF command. Use the prepared firmware if your board appears in the matrix and you want a quick route to a capable voice gadget.

page-agent

alibaba/page-agent · ★ 29k ▲65/7d · TypeScript

Page Agent is an English-documented TypeScript library that places a natural-language agent inside a web page; a Chinese README also exists. It reads a simplified text form of the DOM, asks a configured language model what to do, and can click or fill the interface without screenshots or a headless browser.

Verdict Our Page Agent build and tests passed, but npm audit reported 17 known vulnerabilities, including 3 critical and 9 high, so dependency review comes before embedding it in a signed-in application. Page Agent is a sharp choice for a copilot inside a web product you control because it works from the DOM and needs no remote browser.

agentmemory

rohitg00/agentmemory · ★ 28k ▲248/7d · TypeScript

agentmemory is a local memory service for coding agents, with MCP, hooks, skills, a viewer, and APIs for saving and recalling work across sessions. It supports Claude Code, Codex, Cursor, and other clients, while storing persistent state through a bundled iii engine.

Verdict Our agentmemory run passed build and tests, but its 299-package install used 1,235 MB and npm audit reported 14 known vulnerabilities, including 6 high-severity findings. Try it when cross-session coding context is painful enough to justify a resident service, then restrict capture and inspect recalled records before trusting them.

mastra

mastra-ai/mastra · ★ 28k ▲266/7d · TypeScript

Mastra is a TypeScript framework for building agents, tool-using assistants, graph workflows, and other model-backed applications. It bundles model routing, memory, retrieval, human approval, evaluation, observability, deployment, and Model Context Protocol support so JavaScript teams do not have to assemble each layer separately.

Verdict Our Mastra test command was still running when it hit 900 seconds, after a successful 645-second build, so adopting the full monorepo requires serious CI time. Mastra is a strong fit for TypeScript teams that want agents, controlled workflows, storage, Studio, and MCP in one framework.

qwen-code

QwenLM/qwen-code · ★ 28k ▲157/7d · TypeScript

Qwen Code is an open-source coding agent that runs in a terminal, IDE, desktop shell, daemon, or programmatic SDK. It can use Qwen or other model providers to inspect a repository, edit files, run tools, coordinate subagents, and connect to MCP servers.

Verdict Our Qwen Code run installed 1,937 packages in 348 seconds, used 1,475 MB, and built successfully, but tests were still running at the 900-second cap and npm audit found 9 known vulnerabilities. Try it if provider choice, MCP, and several client surfaces outweigh the size and release pace.

MaxKB

1Panel-dev/MaxKB · ★ 23k ▲58/7d · Python

MaxKB is a web platform for building knowledge-base assistants and multi-step AI agents without writing the whole application from scratch. It combines document ingestion and RAG, a visual workflow builder, model connections, MCP tools, and embeddable chat, with English and Simplified Chinese README documentation.

Verdict Our MaxKB install pulled 224 packages, occupied 1,794 MB, and returned 7 known vulnerability findings, while the repository exposed no test target. Trial v2.10.5-lts when a visual RAG and agent platform can replace custom internal development, but change the default login immediately and validate every document type you depend on.

nuclear

nukeop/nuclear · ★ 18k ▲43/7d · TypeScript

Nuclear is a free desktop music player that searches multiple sources, streams tracks, manages a library, and avoids ads or tracking. Its plugin system supplies music sources and metadata, while an optional MCP server lets AI clients control playback.

Verdict Our Nuclear build completed 8 of 9 tasks, then failed after 67 seconds because `cargo metadata` was unavailable, while all 5 Vitest tests passed. Download the packaged player if you want ad-free desktop streaming and are comfortable selecting source plugins.

mcp-for-beginners

microsoft/mcp-for-beginners · ★ 17k ▲48/7d · Jupyter Notebook

MCP for Beginners is Microsoft's free curriculum for learning the Model Context Protocol through explanations, security guidance, and sample servers and clients. It spans C#, Java, JavaScript, TypeScript, Python, and Rust, then branches into authentication, deployment, Azure integrations, databases, and newer protocol features.

Verdict Our `05-AdvancedTopics` run installed 64 packages and built in 8 seconds, but it exposed no test target, so the measured code path has no suite result. Use this curriculum for breadth, multilingual teaching, and security context, then verify protocol details against the official specification before shipping.

Figma-Context-MCP

GLips/Figma-Context-MCP · ★ 16k ▲60/7d · TypeScript

Framelink MCP for Figma is a Model Context Protocol server that turns Figma file and node data into a smaller description for coding agents. It helps an agent see layout, styling, and selected design structure without relying only on a screenshot.

Verdict Our Framelink MCP run installed 354 packages, built in 18 seconds, and passed its tests in 29 seconds. It is a focused choice for feeding selected Figma structure to Cursor, provided a developer owns the token and verifies every generated interface.

claudian

YishenTu/claudian · ★ 15k ▲91/7d · TypeScript

Claudian puts coding agents such as Claude Code, Codex, Grok, OpenCode, and Pi inside an Obsidian sidebar. It lets an agent read and edit vault files, search notes, run shell commands, use MCP tools, and carry out multi-step work without moving the conversation to a terminal.

Verdict Installing Claudian on our box took 33 seconds but added 862 packages and 906 MB; the build passed in 34 seconds, while tests timed out at 900 seconds. Use it if Obsidian is already your working desk and you understand that the chosen agent can edit files and run commands there.

OpenMetadata

open-metadata/OpenMetadata · ★ 15k ▲59/7d · TypeScript

OpenMetadata is a self-hosted catalog that connects data assets, lineage, quality checks, ownership, governance, and business definitions in one metadata graph. It also exposes that context to assistants through semantic search, SDKs, APIs, and a built-in MCP server.

Verdict Our root run installed 110 packages in 18 seconds but `yarn test` failed in 4 seconds because no test is specified, while the real Docker quickstart requires 6 GiB and 4 vCPUs. OpenMetadata is worth evaluating when catalog context, lineage, governance, and MCP access must live together.

InsForge

InsForge/InsForge · ★ 13k ▲73/7d · TypeScript

InsForge is an open-source backend platform built for applications assembled with coding agents. It combines PostgreSQL, authentication, object storage, edge functions, site deployment, an AI gateway, and an MCP control surface so an agent can inspect and change backend resources while it builds an app.

Verdict Our InsForge run installed 2,284 packages and used 752 MB, then passed its build and tests in 177 seconds combined while npm audit reported 5 vulnerabilities. It is worth a pilot for teams that want agents to operate the same backend primitives they are coding against.

XHS-Downloader

JoeanAmier/XHS-Downloader · ★ 13k ▲59/7d · JavaScript

XHS-Downloader is documented primarily in Simplified Chinese, with a full English README available. It downloads public RedNote posts and metadata through a terminal app, browser userscript, HTTP API, or MCP server, giving people a repeatable way to archive images, videos, and Live Photos.

Verdict Our XHS-Downloader run installed 106 packages in 81 seconds and built in 9 seconds, but it offered no test target, so the code is easy to try and hard to independently regression-check. Use it for hands-on RedNote archiving when its precise media and metadata controls justify maintaining Cookies and watching for site breakage.

ida-pro-mcp

mrexodia/ida-pro-mcp · ★ 12k ▲153/7d · Python

IDA Pro MCP connects an MCP-capable coding agent to IDA Pro so the agent can inspect, annotate, rename, patch, and debug binaries. It supports the IDA interface and headless idalib workers, letting one client manage several named analysis databases.

Verdict Our IDA Pro MCP run installed 35 packages in 22 seconds and built in 8 seconds, but pytest ended with 29 collection or setup errors after 136 tests passed. It is worth trying for an experienced IDA Pro user who will treat agent output as a draft and keep write-capable tools under supervision.

QuantDinger

OpenByteInc/QuantDinger · ★ 12k ▲223/7d · Python

QuantDinger is a self-hosted trading system that connects AI-assisted market research to Python strategies, backtests, paper trading, live execution, and monitoring. It supports crypto exchanges and broker workflows, but the operator still owns the strategy code, risk limits, credentials, and legal responsibility for every order.

Verdict Our QuantDinger run passed 1,415 pytest cases after a 69-second, 159-package install, yet open precision and lot-size reports touch the credibility of backtests. Use it as a serious paper-trading and engineering environment if you can audit those assumptions and operate its service stack.

nginx-ui

0xJacky/nginx-ui · ★ 11k ▲12/7d · Go

Nginx UI is a browser control panel for editing Nginx sites, issuing certificates, reading logs, managing several nodes, and watching server statistics. It also adds an LLM assistant, code completion, and an MCP interface so an AI agent can inspect or change Nginx through the panel.

Verdict Our nginx-ui source install failed in 6 seconds before any build or test could run, so use a pinned release binary or image for evaluation and treat source reproducibility as unresolved. The panel is a good fit for an experienced Nginx operator who wants certificates, backups, nodes, and AI access in one place.

OpenCreator

krillinai/KrillinAI · ★ 11k ▲21/7d · TypeScript

OpenCreator, formerly KrillinAI, is a local workspace that uses Codex to help make and edit videos, images, voiceovers, subtitles, and other creative files. It combines an agent conversation with visual tools, so a creator can move between instructions and hands-on adjustments without moving the job to another app. The canonical README is in English and links to 10 maintained translations.

Verdict Our 2026-08-22 KrillinAI run installed 239 Go packages and built successfully, but 3 of 15 package results failed, and the live repository has since become TypeScript-based OpenCreator v3.1.0. Try it if you already trust Codex with local projects and want one interface for agent work plus media production.

opensre

Tracer-Cloud/opensre · ★ 11k ▲60/7d · Python

OpenSRE is a self-hosted framework for AI agents that investigate production incidents across logs, metrics, traces, deployments, and runbooks. It gives operators an interactive shell, a headless command, and hosted deployment paths for collecting evidence and producing a root-cause report.

Verdict Our OpenSRE run installed 201 packages and passed 15,768 tests, but 23 failures kept the suite red after 823 seconds, so this is a serious evaluation target rather than a safe default for production response. Try it if your SRE team needs one agent across many data sources and will constrain its permissions.

mcp-use

mcp-use/mcp-use · ★ 11k ▲42/7d · TypeScript

mcp-use is a TypeScript and Python toolkit for connecting AI applications to Model Context Protocol servers, the small programs that expose tools and data to models. It can create MCP servers and interactive views, connect to existing servers, or run a LangChain-based Python agent across several of them.

Verdict Our mcp-use Python build succeeded and pip-audit found 0 known vulnerabilities, but only 159 of 337 pytest cases passed while 135 failed and 43 hit errors, so adopt it behind your own tool-level integration tests. The TypeScript side is a strong trial for teams that want MCP tools and interactive views in one scaffold.

aichat

sigoden/aichat · ★ 10k ▲17/7d · Rust

AIChat is a Rust command-line client for talking to hosted and local language models from a shell, an interactive REPL, or a small HTTP server. It also handles files, saved roles and sessions, document retrieval, function calls, MCP tools, and agent definitions so one executable can replace several model-specific clients.

Verdict Our AIChat run installed 425 packages and passed all 21 tests after a 152-second build, so the checked-out Rust client is easy to trust for a personal trial. Use it if you want one terminal interface across many model providers and are comfortable owning keys, local config, and tool permissions.

mcp

awslabs/mcp · ★ 9.7k ▲19/7d · Python

awslabs/mcp is a collection of Model Context Protocol servers for AWS documentation, infrastructure, databases, observability, cost, messaging, and other services. It lets MCP clients such as Claude Code and Cursor call focused AWS tools, but AWS now recommends Agent Toolkit for AWS for production software.

Verdict Our run built the 71 MB repository's Docusaurus site in 59 seconds, but it ran no server tests and npm audit found 31 advisories, so that pass says nothing about the operational safety of each MCP package. Use awslabs/mcp to evaluate focused AWS integrations or study their code.

feynman

advaitpaliwal/feynman · ★ 9.6k ▲364/7d · TypeScript

Feynman is an MIT-licensed AI research agent that searches papers and the web, ranks literature, checks claims against code, and produces cited research briefs. It runs as a terminal app or local science workbench, with separate research skills available for Codex and repository-based agents.

Verdict Our Feynman run installed 985 packages and used 1,329 MB, then its tests failed after 699 seconds, so this is a serious research workspace rather than a lightweight paper-search utility. The 27-second build and zero-advisory audit support a trial, but the failing assertion and rapid release corrections argue for a pinned version and a local acceptance suite.

gsd-core

open-gsd/gsd-core · ★ 9.5k ▲199/7d · JavaScript

GSD Core is a workflow system that guides AI coding agents through discussing, planning, building, checking, and shipping a software project. It stores decisions and progress in project files so a fresh agent can continue the work without depending on one long chat history.

Verdict Our GSD Core run installed 416 packages and built successfully, but its test process was still alive at the 900-second cap after reporting 1,733 passes and zero failures. Adopt it for substantial agent-led projects when durable planning files and repeatable phase checks justify a large workflow layer.

openwhispr

OpenWhispr/openwhispr · ★ 8.2k ▲373/7d · JavaScript

OpenWhispr is a desktop voice-to-text app for dictation, meeting transcription, searchable notes, and spoken commands to AI models. It can keep transcription and language-model work on the computer, or connect to hosted providers when speed, sync, or collaboration matters more than local processing.

Verdict Our OpenWhispr install used 1,374 MB, its package build exceeded 900 seconds, and its tests ended on a rollback failure, so source adoption needs more patience than the quick start implies. The released app is worth trying for people who genuinely need dictation, meetings, and local AI in one desktop client.

Graft

trailhq/Graft · ★ 8.1k ▲1.6k/7d · TypeScript

Graft builds a local map of a codebase so coding agents can find symbols, relationships, and likely files without rereading the repository from scratch. Its structural graph works without a model key, while an optional deep pass asks your chosen model provider to write richer explanations.

Verdict Our Graft run installed 86 packages, passed all 938 tests, and reported 0 known vulnerabilities, making it an unusually low-friction trial for teams whose coding agents keep rediscovering the same code. Use the structural graph first, inspect `graft init --dry-run`, and add the model-backed deep pass only if its extra explanations justify sending source-derived content to your provider.

osaurus

osaurus-ai/osaurus · ★ 7.9k ▲107/7d · Swift

Osaurus is an English-documented, native macOS application for running local or cloud AI models with agents, memory, tools, and an OpenAI-compatible server. It keeps the surrounding agent state on the Mac and can also act as both an MCP server and a client for remote MCP providers.

Verdict Our sandbox passed 1 of 1 tests for the Go helper in 31 seconds, but it did not build the 848,110-line Swift-centered Mac application that users actually install. Osaurus is worth trying for an Apple Silicon owner who wants local inference, agents, MCP, and automation in one native app, especially if macOS 26 enables the Linux VM sandbox.

code2prompt

mufeedvh/code2prompt · ★ 7.7k ▲18/7d · Rust

Code2Prompt collects selected files, a directory tree, and optional Git context into one document that you can give to a language model. It comes as a Rust command-line app and library, an interactive terminal interface, and Python bindings, while the README also advertises a local MCP service.

Verdict Our Code2Prompt build succeeded in 282 seconds and all 336 tests passed in 96 seconds, so the implementation earns a trial while the source compile favors packaged binaries for occasional use. Choose it for a reviewable repository snapshot with strong filtering and formatting controls.

Soup

MakazhanAlpamys/Soup · ★ 6.6k ▲695/7d · Python

Soup is a Python command-line toolkit for fine-tuning, evaluating, and packaging language models from one YAML recipe. It tries to replace a pile of training scripts and infrastructure chores with a guided workflow that can run locally, including an experimental layer-streaming mode for GPUs with very little memory.

Verdict Our Soup install used 5,370 MB and its test suite reached only 31% before a 900-second timeout, so contributors should budget for a large environment and a long verification loop. Try it for a guided local path from dataset to evaluated adapter, starting with an ordinary LoRA job.

zcf

UfoMiao/zcf · ★ 6.1k ▲2/7d · TypeScript

ZCF is an interactive installer and configuration manager for Claude Code and Codex. It can install workflows, set model providers, configure Claude Code Router and MCP services, and apply language or assistant preferences from one command instead of asking users to edit several files by hand.

Verdict Our ZCF run installed 899 packages in 27 seconds, then passed its 13-second build and 50-second test command, so the code is easy to verify but the resulting 319 MB toolchain is not truly zero configuration. Use it when repeatable Claude Code and Codex setup saves more time than reviewing the files it changes.

openpencil

ZSeven-W/openpencil · ★ 5.9k ▲49/7d · Rust

OpenPencil is a Rust vector-design application that lets people and AI agents create and edit structured canvas documents. Its `.op` files are JSON, and the same project includes desktop and web editors, a CLI, an MCP server, collaboration, and code export.

Verdict Our OpenPencil build took 448 seconds and all 547 tests passed, making commit 9c81077 a credible engineering base for teams willing to test prerelease desktop behavior themselves. Use it when JSON design files, MCP control, and local ownership solve a concrete workflow problem.

openmed

maziyarpanahi/openmed · ★ 5.3k ▲53/7d · Python

OpenMed is a local-first toolkit for finding medical entities and personal identifiers in clinical text, then masking or replacing sensitive details. It supplies Python, service, mobile, browser, MCP, and coding-agent paths, but deployment owners still have to validate the models and clinical output for their own data.

Verdict Our OpenMed run installed 150 packages and 941 MB, then showed a failure marker before tests hit the 900-second cap, so production adoption needs more validation than the quick examples imply. The 16-second build and 0 known audited vulnerabilities make a controlled trial reasonable for teams that need local clinical text processing across several runtimes.

eve

vercel/eve · ★ 5.2k · TypeScript

Eve is a TypeScript framework for agents that may need to pause, resume, run on a schedule, call tools, or talk through web and messaging channels. You build an agent by placing instructions, tools, skills, connections, channels, schedules, and subagents in named filesystem locations, then Eve compiles and runs that layout.

Verdict Our Eve run installed 2,573 packages and used 2,815 MB, then 8,673 tests passed and 1 was skipped with no failures. That is strong evidence for the repository at commit 32aca9b, but it does not make Eve lightweight or finished.

kagent

kagent-dev/kagent · ★ 3.7k ▲51/7d · Go

Kagent is a Kubernetes-native framework for defining, running, and observing AI agents as cluster resources. It lets platform teams manage agent prompts, models, MCP tools, and deployments with YAML and `kubectl` instead of building a separate orchestration service from scratch.

Verdict Our Kagent run built successfully, but 7 of 73 tests failed after a 162-second test step. It is worth an isolated-cluster trial when agent sprawl has become a Kubernetes platform problem.

kandev

kdlbs/kandev · ★ 795 ▲29/7d · Go

Kandev is a self-hosted development control plane that turns coding-agent work into Kanban tasks, isolated worktrees, review steps, and pull requests. It can run many agent CLIs through local processes, Docker, SSH, or cloud executors, then place their conversations, diffs, terminals, files, and previews in one workspace.

Verdict Our Kandev install took 48 seconds and 978 MB, but the build failed after 7 seconds because the desktop Tauri step could not run `cargo metadata`; no tests ran. Use a packaged release if its Kanban, worktree, review, and multi-executor model matches how your team already supervises several agents.

ECC

affaan-m/ECC · ★ 259k ▲5.0k/7d · JavaScript

ECC is a large set of skills, specialist agents, rules, hooks, memory tools, and installers for coding assistants, with Claude Code as its best-supported host. It tries to make planning, tests, review, security checks, and handoffs repeatable instead of asking developers to restate their process in every prompt.

Verdict ECC makes sense for a heavy Claude Code user who wants one maintained opinion about planning, verification, memory, and agent roles. The successful test run is reassuring, but the product is too broad and too dependent on host-specific hooks to install casually.

hermes-agent

NousResearch/hermes-agent · ★ 246k ▲2.4k/7d · Python

Hermes Agent is an English-documented personal AI agent that runs in a terminal or behind messaging gateways for Telegram, Discord, Slack, WhatsApp, Signal, and email. It can call shell and file tools, schedule jobs, remember past work, create reusable skills, delegate tasks, and switch among hosted or self-run model providers; Chinese, Urdu, and Spanish README translations also exist.

Verdict Hermes is compelling for an experienced self-hoster who wants one agent to persist across a terminal, a server, and messaging apps. Its provider choice, memory, skills, scheduling, and remote backends cover far more than a chat wrapper, but that scope creates a large trust and maintenance burden.

AutoGPT

Significant-Gravitas/AutoGPT · ★ 187k ▲131/7d · Python

AutoGPT is now a platform for building and running agent workflows through plain-English setup or a visual block editor, with schedules, triggers, integrations, costs, and approvals around each run. The same repository also retains AutoGPT Classic, the original autonomous-agent experiment, as unsupported educational code.

Verdict Our AutoGPT Classic run passed 277 of 380 tests, failed 65, hit 38 setup or collection errors, and exposed 55 known vulnerabilities, matching the maintainers' warning that Classic is unsupported. Do not deploy Classic; use it only to study the original experiment.

prompts.chat

f/prompts.chat · ★ 170k ▲626/7d · HTML

prompts.chat is a public prompt catalog and a self-hostable web application for collecting, revising, searching, and sharing prompts. Teams can run a branded private library, while individual users can browse the hosted catalog or reach it from Claude Code and other MCP clients.

Verdict Our prompts.chat install stopped after 164 seconds because `DATABASE_URL` was required during postinstall, so the documented manual order does not work in a blank environment. The hosted catalog and Claude Code plugin are easy ways to sample the project without operating it.

cc-switch

farion1231/cc-switch · ★ 133k ▲1.3k/7d · Rust

CC Switch is a desktop manager for provider accounts, model routes, MCP servers, skills, and local configuration across Claude Code, Claude Desktop, Codex, Gemini CLI, Grok Build, OpenCode, OpenClaw, and Hermes. Its default README is in English, with Chinese, Japanese, and German versions also available.

Verdict Our CC Switch install failed after 11 seconds because its pnpm configuration set two conflicting built-dependency policies, so we could not reach a build or test result. The desktop release may still save serious time for people juggling 8 supported agents, but configuration takeover deserves a backup drill and a trial account before daily use.

open-design

nexu-io/open-design · ★ 96k ▲1.2k/7d

OpenDesign is a local-first design workspace that asks coding agents to create real HTML, documents, slides, images, and motion graphics from prompts and reusable design systems. It connects Claude Code, Codex, DeepSeek Harness, and many other command-line agents to a desktop or web studio where generated files can be previewed, critiqued, edited, and exported.

Verdict Our OpenDesign install pulled 1,293 packages and used 2,644 MB, then exposed no root build or test target. Try it if you already trust coding agents with project files and want brand rules, previews, and export in one workspace.

claude-mem

thedotmack/claude-mem · ★ 94k ▲395/7d · JavaScript

Claude-Mem gives coding agents a searchable memory that persists across sessions. It watches agent activity, compresses the useful parts with a selected AI provider, stores the result, and injects relevant project history when later sessions begin.

Verdict Our Claude-Mem run installed 799 packages and used 995 MB, then its tests stopped after 8 seconds because Bun was absent. Try it when repeated project rediscovery costs more than operating a local worker and auditing captured data.

ragflow

infiniflow/ragflow · ★ 91k ▲366/7d · Go

RAGFlow is a self-hosted system for turning documents and connected data sources into searchable knowledge bases, cited chat answers, and agent workflows. It bundles parsing, chunk review, retrieval, reranking, model connections, APIs, and a visual interface so a team does not have to assemble each layer separately.

Verdict RAGFlow is a strong candidate when document parsing and chunk inspection are central to the product, and the team wants a usable interface as well as APIs. It is an application stack, not a drop-in RAG package, so budget for operations, migration tests, provider configuration, and document-specific acceptance tests.

deer-flow

bytedance/deer-flow · ★ 82k ▲418/7d · Python

DeerFlow is an open-source agent application that can research, write code, create files, remember user context, and split work among sub-agents. It combines a web interface, model providers, tools, skills, MCP connections, and isolated execution so longer tasks can run with more structure than a single chat response.

Verdict Our DeerFlow backend installed 33 packages and built in 16 seconds, but pytest ran 0 tests and stopped after 200 collection or setup errors, so commit 13f0a7f is a trial candidate rather than a verified deployment. Use DeerFlow when you need a full agent application with model choice, memory, skills, sub-agents, and sandboxes, and you have an operator who understands its privilege.

Scrapling

D4Vinci/Scrapling · ★ 81k ▲1.8k/7d · Python

Scrapling is a Python scraping framework that combines HTML parsing, HTTP fetching, browser automation, adaptive selectors, spiders, and an MCP server. It is meant for jobs that start with one page but may grow into concurrent crawls with sessions, proxies, pause and resume, or agent access.

Verdict Our Scrapling run installed in 24 seconds and passed 258 tests, but 5 failed and 110 more hit setup errors, with the log showing `curl_cffi` was missing. Try it when one Python project genuinely needs HTTP requests, browsers, adaptive parsing, spiders, and MCP access under one API.

ai-agents-for-beginners

microsoft/ai-agents-for-beginners · ★ 75k ▲506/7d · Jupyter Notebook

AI Agents for Beginners is an English-language Microsoft course, with more than 50 automated translations, that teaches agent design through 18 lessons and Jupyter notebooks. It covers tools, RAG, planning, multiple agents, memory, MCP and other protocols, browser use, deployment, local models, and security, with current examples centered on Microsoft Agent Framework and Foundry.

Verdict Our run installed 172 packages and 1,301 MB in 207 seconds, then built in 18 seconds but failed before any test ran because `demo_tool_agent` was missing during collection. Use the course if Microsoft Foundry is a realistic target and you want 18 guided lessons spanning agent patterns through deployment and security.

TrendRadar

sansan0/TrendRadar · ★ 62k ▲151/7d · Python

TrendRadar is a Chinese-first news and trend monitor, with an English README and English documentation available. It collects hot lists and RSS feeds, filters the results by keywords or AI-defined interests, keeps ranking history, and sends reports to chat, email, push, or webhook destinations.

Verdict Our TrendRadar install pulled 119 packages, and pip-audit found 25 known vulnerabilities, so adoption requires active dependency review from day one. It is a good fit for a Chinese-speaking individual or small team whose sources already match NewsNow and who wants one configurable path from collection to delivery.

ai-engineering-from-scratch

rohitg00/ai-engineering-from-scratch · ★ 55k ▲1.5k/7d · Python

AI Engineering From Scratch is a free curriculum that teaches the math, code, model systems, agents, MCP, and production work behind modern AI through runnable lessons. English is canonical; translated landing pages are committed in 12 languages, while translated lesson pages are machine-produced on a separate branch.

Verdict Our install consumed 5,609 MB across 204 packages and the repository exposed no unified test target, so this is a serious study archive rather than a tidy course package. Use it if you want a self-directed route from linear algebra to agents and can verify each lesson as you go.

cherry-studio

CherryHQ/cherry-studio · ★ 52k ▲223/7d · TypeScript

Cherry Studio is a desktop client for chatting with hosted and local language models on Windows, macOS, and Linux. It puts multiple providers, custom assistants, documents, translation, search, backups, and MCP tools in one graphical app instead of making users manage separate chat tabs.

Verdict Our Cherry Studio install consumed 3,552 MB, the build failed on a missing DSH bridge module, and tests were still running at the 900-second cap. Downloaded releases may still suit an individual who wants many model providers in a polished desktop client, but this commit is a costly starting point for contributors.

CopilotKit

CopilotKit/CopilotKit · ★ 37k ▲98/7d · TypeScript

CopilotKit is a TypeScript toolkit for putting an agent inside a web, mobile, Slack, or Microsoft Teams interface. It provides chat components, shared state, tool rendering, generated UI, and approval flows so a product team does not have to build the agent-facing layer from scratch.

Verdict Our CopilotKit install took 275 seconds and 3,935 MB, then both the build and test commands exited 130, so adopting it means accepting a large, fast-moving platform rather than adding a small chat widget. It is a serious candidate for React products that need shared state, rendered tools, and human approval in the interface.

PDFMathTranslate

PDFMathTranslate/PDFMathTranslate · ★ 37k ▲172/7d · Python

PDFMathTranslate turns a scientific PDF into translated and bilingual PDFs while trying to preserve formulas, charts, annotations, and page layout. It can run through a command line, browser interface, Docker, Python or HTTP APIs, and MCP, with English documentation plus Chinese, Japanese, and Korean README translations.

Verdict Our PDFMathTranslate run built in 13 seconds, but tests ended with 3 failures and 5 errors while pip-audit found 45 vulnerabilities, so this checkout is suitable for supervised evaluation rather than blind document processing. Use it when preserving scientific layout is valuable and someone will compare the output with the source.

next-ai-draw-io

DayuanJiang/next-ai-draw-io · ★ 36k ▲171/7d · TypeScript

Next AI Draw.io is a web and desktop application that turns natural-language requests, documents, and images into editable draw.io diagrams. It lets people ask a model to create or revise diagram XML while keeping the result inside the familiar draw.io editor.

Verdict Our install consumed 2,041 MB and npm audit found 35 known vulnerabilities, including 1 critical and 25 high, even though the 48-second build and 9-second tests passed. Next AI Draw.io is worth trying locally when editable draw.io output is more useful than an image, especially through its MCP server.

nacos

alibaba/nacos · ★ 33k ▲17/7d · Java

Nacos is a control-plane service for dynamic configuration, service discovery, health checks, DNS-based routing, and service metadata. Its 3.x line also manages AI resources such as MCP servers, agent specifications, A2A entries, prompts, and skill packages, giving AI platforms a registry beside the older microservice features.

Verdict Our Nacos console build completed in 48 seconds, but it installed 1,399 packages and npm audit found 99 vulnerabilities, including 40 critical or high. Existing Nacos shops should evaluate the 3.x AI registry because it can keep MCP, A2A, skills, and ordinary service metadata in one control plane.

oh-my-pi

can1357/oh-my-pi · ★ 31k ▲912/7d · TypeScript

Oh My Pi, invoked as `omp`, is a terminal coding agent that can edit code, run shells, use language servers and debuggers, browse the web, coordinate subagents, and remember a project. It is a heavily extended fork of Pi for developers who want one configurable agent interface across many model providers.

Verdict Our Oh My Pi checkout pulled 412 packages and 1,517 MB, then its native build failed after 225 seconds and all 12 test files errored, so source contributors should use the pinned Nix environment before judging the code. As an installed CLI, `omp` is compelling for developers who genuinely want model routing, native code tools, browser control, subagents, and memory under one roof.

cognee

topoteretes/cognee · ★ 31k ▲106/7d · Python

Cognee is a memory layer for AI agents that turns documents and conversation history into searchable vectors and a knowledge graph. It gives applications `remember`, `recall`, `forget`, and `improve` operations, with local, self-hosted, cloud, MCP, and Claude Code paths.

Verdict Our Cognee install consumed 1,126 MB and its tests were still unfinished at the 900-second cap, so adopting it means budgeting for a large, moving system rather than a small memory helper. It is worth testing when persistent agent memory, document ingestion, and graph retrieval belong in one service and your team can measure answer quality against its own corpus.

composio

ComposioHQ/composio · ★ 30k ▲94/7d · TypeScript

Composio is a hosted action layer for AI agents, with SDKs that let an agent find tools, connect a user's account, and call outside services. It saves developers from building and maintaining authentication plus API wrappers for each app their agent uses.

Verdict Our Composio checkout installed 910 packages and occupied 2,280 MB, then both build and tests failed because Bun was missing. Use the hosted product when a broad connector catalog and per-user authentication remove more work than an external control plane adds.

picoclaw

sipeed/picoclaw · ★ 30k ▲44/7d · Go

PicoClaw is a Go-based personal AI assistant that connects language models to a terminal, browser interface, chat apps, schedules, local tools, and MCP servers. It targets small Linux boards and old Android phones, giving one low-cost device an always-available agent without requiring a desktop-sized host.

Verdict Our PicoClaw run passed 127 of 134 tests but failed 7 after 766 seconds, matching the project's own warning that releases before v1.0 are not for production. It is an interesting personal-lab choice when small hardware, many provider options, chat channels, and MCP matter enough to justify careful isolation.

gpt-researcher

assafelovic/gpt-researcher · ★ 29k ▲111/7d · Python

GPT Researcher is a Python research agent that searches the web or local documents, keeps track of sources, and turns the material into a cited report. It can run as a library, API, web application, Claude Code skill, or client of MCP data sources.

Verdict Our GPT Researcher run installed 230 packages and passed 398 of 412 tests, but the build and full test step both ended with exit code 1. It remains a sensible base for developers who need cited research across web pages and local files and are prepared to inspect the failures and the generated report.

kilocode

Kilo-Org/kilocode · ★ 27k ▲86/7d · TypeScript

Kilo Code is an open-source AI coding agent that works in VS Code, JetBrains IDEs, the terminal, and a web-hosted cloud agent. It handles code generation, repository questions, debugging, review, autocomplete, and command execution, so developers do not need a different assistant for each interface.

Verdict Our install completed in 80 seconds, but it pulled 1,964 packages and the root test command failed by design after 18 seconds. Kilo Code is worth trying if you genuinely want the same agent across VS Code, JetBrains, CLI, cloud, and automated review, especially when model choice matters.

gin-vue-admin

flipped-aurora/gin-vue-admin · ★ 25k ▲17/7d · Go

Gin-Vue-Admin is a Chinese-first full-stack admin starter; an English README exists, but its setup and branch guidance are older than the Chinese documentation. It combines a Gin backend, Vue interface, permissions, code generation, MCP tools, and AI project guidance for teams building internal business software.

Verdict Our Gin-Vue-Admin run installed 532 packages and compiled in 101 seconds, while 5 of 27 Go package results failed, leaving a usable build and a red test baseline. It is worth a trial for a Chinese-speaking Go and Vue team that wants an opinionated admin product with MCP-assisted generation.

context-mode

mksglu/context-mode · ★ 23k ▲1.7k/7d · TypeScript

Context Mode is a local MCP server and coding-agent add-on that keeps bulky tool output out of the conversation window. It runs analysis in child processes, stores searchable material and session events in SQLite, then returns only selected output. Hook-capable clients can also restore working state after compaction.

Verdict Our 201-package Context Mode install built in 12 seconds, but the test command was still running at our 900-second cap, so installation is easier to trust than full verification. Use it when logs, web pages, and repeated file reads are consuming an agent's context and your client supports its hooks.

n8n-mcp

czlonkowski/n8n-mcp · ★ 23k ▲42/7d · TypeScript

n8n-MCP is a Model Context Protocol server that lets AI assistants look up n8n nodes, inspect templates, validate workflows, and manage an n8n instance. It solves the tedious part of giving an assistant accurate node schemas and operations instead of asking it to invent workflow JSON from memory.

Verdict Our n8n-MCP run installed 2,761 packages and passed 6,387 of 6,660 tests, while 220 failed and npm audit reported 133 known vulnerabilities. That makes it useful for assisted discovery and drafting, but a poor candidate for unsupervised production writes.

cua

trycua/cua · ★ 23k ▲264/7d · HTML

Cua is a collection of tools for agents that operate graphical computers: background desktop drivers, cross-OS sandboxes, evaluation environments, and macOS virtual machines. It gives developers APIs and an MCP server for seeing screens, clicking, typing, running shell commands, and collecting trajectories across Linux, macOS, Windows, and Android.

Verdict Our Cua root run installed only 2 packages in 11 seconds, then found no build or test target, so it validated a thin pnpm surface rather than this 4,255-file cross-OS platform. Cua is worth a focused trial when an agent must control native applications or when research needs repeatable desktop environments.

screenpipe

screenpipe/screenpipe · ★ 22k ▲83/7d · Rust

Screenpipe records screen activity and audio into a searchable local history, then lets AI agents query that history through an API and MCP. It uses accessibility data where available, falls back to OCR, stores records in SQLite, and can run scheduled agent prompts called pipes.

Verdict Our Screenpipe install pulled 1,144 packages, then the 846-second build and 278-second test run both stopped because `libpipewire-0.3` was unavailable. Trial the signed app if local searchable memory and MCP access justify recording your work, but review telemetry, retention, and pipe permissions before the recorder starts.

Auto-claude-code-research-in-sleep

wanshuiyin/Auto-claude-code-research-in-sleep · ★ 16k ▲276/7d · Python

ARIS is an English-first, partly bilingual collection of agent skills and a standalone CLI for running machine-learning research loops. A separate Chinese README exists, while the main English documentation explains how Claude Code, Codex, and other agents can search literature, propose experiments, review results, and draft papers.

Verdict Our Feishu bridge run installed 43 packages and built in 33 seconds combined, but pytest collected 0 tests and exited 5. ARIS is worth studying or piloting if you already have an ML project, two suitable model routes, and a human who will audit what ran overnight.

superset

superset-sh/superset · ★ 14k ▲258/7d · TypeScript

Superset is a desktop workspace for running several command-line coding agents against separate Git worktrees. It combines terminals, agent status, diffs, previews, scheduled jobs, remote hosts, a CLI, an SDK, and an MCP server in one interface.

Verdict Our Superset checkout installed 3008 packages and occupied 5661 MB, then failed both its 244-second build and 49-second test step. The released macOS app is still worth trying for developers already juggling several paid CLI agents, because worktree isolation and one review surface solve a real coordination problem.

BrowserOS

browseros-ai/BrowserOS · ★ 14k ▲84/7d · TypeScript

BrowserOS contains two Chromium-based browsers: BrowserOS neo gives coding agents a separate logged-in browser, while BrowserOS gives a person a daily browser with a built-in agent. Both run locally, and their MCP interfaces let tools such as Claude Code and Codex control browser tabs and web tasks.

Verdict Our BrowserOS package install took 35 seconds and pip-audit found 0 known vulnerabilities, but the build exited 1 after 6 seconds and no tests ran. BrowserOS neo is worth a supervised trial when an agent needs persistent login state and session replay on a local machine.

holaOS

holaboss-ai/holaOS · ★ 11k ▲103/7d · TypeScript

holaOS is an Electron workspace that puts AI agents beside browser pages, connected apps, chat services, local files, and shared memory. It gives Claude Code, Codex, its built-in agent, and MCP tools one desktop environment instead of keeping each tool in a separate chat or terminal.

Verdict Our holaOS install consumed 3,425 MB for 2,035 packages, then built in 73 seconds while the workspace test command still exited 1, so this is a substantial desktop platform with unfinished verification edges. It is worth a controlled trial for teams that genuinely want Claude Code, Codex, MCP, connected apps, and shared local memory in one interface.

claurst

Kuberwastaken/claurst · ★ 10k ▲12/7d · Rust

Claurst is a GPL-licensed terminal coding agent written in Rust, with a TUI, file and shell tools, subagents, plugins, memory, MCP support, and multiple model providers. It began as a clean-room reimplementation of Claude Code behavior and now also runs through Agent Client Protocol editors such as Zed.

Verdict Our npm-path install used 35 MB and reported 0 vulnerabilities, but it exposed no build or test target and therefore supplied no evidence about the Rust agent. Claurst is an appealing beta for developers who want Claude Code-style operation with provider freedom, MCP, and ACP in one native binary.

stitch-skills

google-labs-code/stitch-skills · ★ 8.3k ▲47/7d · TypeScript

Stitch Skills is a collection of agent skills and plugins that connects coding agents to Google Stitch for design generation, design-system work, code conversion, and related build tasks. It solves the awkward handoff between an AI design workspace and the agents developers already use to write application code.

Verdict Use Stitch Skills if Google Stitch is already part of your workflow and you want Codex or another supported agent to carry designs into code with repeatable instructions. The breadth is useful and the repository looks active, but the MCP prerequisite and uneven installation paths make it a poor casual trial.

unstract

Zipstack/unstract · ★ 7.2k ▲11/7d · Python

Unstract turns PDFs, scans, office documents, and images into structured JSON using prompts and language models. It packages that extraction work as a visual studio, REST API, ETL pipelines, connectors, and an MCP server instead of leaving teams to assemble each part themselves.

Verdict Our Unstract build passed in 14 seconds, but its test command collected no runnable tests because 2 suites stopped on a missing `yaml` module. The platform is worth a controlled trial when document extraction must become an API or ETL product and your team already operates Docker services and model credentials.

agent-governance-toolkit

microsoft/agent-governance-toolkit · ★ 6.3k ▲40/7d · Python

Microsoft's Agent Governance Toolkit adds policy checks, identity, audit records, sandbox controls, and operational limits around AI agents. It wraps tool calls and agent actions in ordinary application code so an allow, deny, or approval decision can happen before the action reaches an external service.

Verdict Our measured Antigravity CLI subproject installed 7 packages and passed 22 tests, but npm audit found 2 high-severity vulnerabilities, so even the small path needs dependency review before use. Trial the toolkit when an agent can cause real external effects and your team can verify that interception actually fires on every supported host.

byterover-cli

campfirein/byterover-cli · ★ 5.0k ▲3/7d · TypeScript

ByteRover CLI gives coding agents a persistent project memory made from files and facts you curate. It runs as a local command-line tool and web interface, can share memory through ByteRover Cloud, and exposes an MCP server for compatible agents.

Verdict Our ByteRover run passed all 8,910 tests, but installing it pulled 1,673 packages and npm audit found 46 known vulnerabilities, including 16 high-severity findings. Try it for reviewed, portable coding-agent memory if your team will inspect each curation result and can absorb its 1,006 MB dependency footprint.

open-science

aipoch/open-science · ★ 4.2k ▲173/7d · TypeScript

AIPOCH Open Science is a local-first desktop workbench where researchers can ask AI agents to search, run Python or R, use scientific data connectors, and produce files with provenance. It keeps projects, conversations, tool activity, notebooks, and generated artifacts together instead of scattering a research task across a chat app and separate scripts.

Verdict Our Open Science install pulled 1,682 packages and 2,220 MB, while its test command still had not finished after 900 seconds, so adopting it from source is a serious workstation commitment. It is worth a trial for an individual researcher who values visible tool activity, local project records, and scientific connectors.

Hyper-Extract

yifanfeng97/Hyper-Extract · ★ 3.9k ▲19/7d · Python

Hyper-Extract is a Python CLI and library that turns unstructured documents into typed collections, graphs, hypergraphs, and searchable knowledge stores with an LLM. It is aimed at developers who want reusable extraction templates and local artifacts instead of writing a separate prompt and parser for every document type.

Verdict Our Hyper-Extract run installed 118 packages and built successfully, but 4 of 318 pytest cases failed because `langchain_anthropic` could not be imported. Use it when the 80-plus templates and several graph shapes save more work than maintaining provider configuration and validating model output.

openscience

synthetic-sciences/openscience · ★ 3.6k ▲43/7d · TypeScript

OpenScience is a local browser workbench where an AI agent can search scientific sources, write code, run experiments, and assemble results in one project. It combines a research chat, files, terminal access, scientific viewers, provider routing, and a large library of domain instructions.

Verdict Our OpenScience test command was still running at the 900-second cap after an 81-second successful build, so adopting this workbench means budgeting for a large and slow-moving verification surface. It is worth a controlled trial for a technical lab that wants literature, code, compute, and provenance in one local interface.

buildwithclaude

davepoon/buildwithclaude · ★ 3.5k ▲35/7d · Python

Build with Claude is a directory and installable marketplace for Claude Code agents, commands, hooks, skills, plugins, and links to MCP servers. It helps developers find reusable Claude Code extensions without hunting through scattered repositories, while its web interface provides search and filtering.

Verdict Build with Claude is useful as a search layer and a starting point for teams already committed to Claude Code. Do not mistake catalog size for curation depth: install selected entries after reading their permissions and code, especially hooks.

locally-uncensored

PurpleDoubleD/locally-uncensored · ★ 1.7k ▲178/7d · TypeScript

Locally Uncensored is a desktop app for Windows and Linux that puts local AI chat, coding, image generation, and video generation in one interface. It tries to replace several model front ends with one installer, while still letting experienced users connect their existing local engines or optional cloud accounts.

Verdict Our Locally Uncensored run installed 477 packages into 609 MB and found 5 high-severity advisories, so security-conscious teams should trial it on an isolated workstation first. It is worth trying when one local desktop needs to cover chat, coding, images, and video.

doop

kgoedecke/doop · ★ 696 · TypeScript

Doop is an open-source multiplayer design canvas where people and software agents can work on the same interface designs in real time. It solves the awkward handoff between an AI generating markup and a designer repeatedly refreshing a separate preview.

Verdict Our build completed in 13 seconds, and 350 of 361 tests passed while 11 were skipped. Doop is worth trying if live, inspectable collaboration between designers and MCP-connected agents is the actual requirement, not just an AI label on a drawing tool.

autonomous-os

autonomous-ai/autonomous-os · ★ 337 ▲20/7d · Python

Autonomous OS is a software stack that connects robot hardware, voice services, safety rules, agent runtimes, and installable skills. It gives supported robots a common way to see, hear, speak, move, and hand tasks to engines such as Hermes, Codex, or Claude Code without putting servo control inside the language model.

Verdict Our Autonomous OS checkout installed 65 packages and passed all 34 tests, which makes the Go core unusually easy to evaluate for a robot stack of this size. Use it for an experimental embodied agent when its supported bodies and replaceable runtimes match your design.

oh-my-openagent

code-yeongyu/oh-my-openagent · ★ 69k ▲220/7d · TypeScript

oh-my-openagent is an agent orchestration layer for OpenCode, Codex CLI, and a beta standalone runtime. It installs specialist agents, hooks, skills, MCP tools, planning modes, and persistent work loops so a coding agent can split large jobs across several workers.

Verdict Our oh-my-openagent install consumed 3,058 MB and its test suite was still running at 900 seconds, so this is a serious harness commitment rather than a casual plugin. Try it if you already use OpenCode or Codex, need persistent multi-agent work, and will set hard limits on time, tokens, permissions, and fan-out.

OmniRoute

diegosouzapw/OmniRoute · ★ 67k ▲3.6k/7d · TypeScript

OmniRoute is a self-hosted AI gateway that gives coding tools and other clients one OpenAI-compatible endpoint for many model providers. It stores provider connections, routes requests by policy, retries or falls back when targets fail, and adds a dashboard for quotas, usage, keys, compression, and gateway controls.

Verdict Our OmniRoute install consumed 5,276 MB, its production build failed after 64 seconds, and its tests exceeded 900 seconds, so the measured commit is too heavy and uncertain for an unqualified production recommendation. It is worth a contained trial for a power user juggling several model providers and coding clients through one endpoint.

JeecgBoot

jeecgboot/JeecgBoot · ★ 48k ▲141/7d · Java

JeecgBoot is a Chinese-first, self-hosted low-code platform for building Java business systems, and a substantial English README is available. It combines generated Spring Boot and Vue code with forms, workflows, reports, dashboards, permissions, and an AI application layer for teams that would otherwise assemble those pieces themselves.

Verdict Our JeecgBoot front-end install failed after 9 seconds, so the 3,497-file platform is a poor choice for teams seeking a quick, low-risk starter. It is worth a controlled trial for Java shops that genuinely need its unusual mix of editable code generation, workflow, reporting, permissions, and AI features.

ai-agent-book

bojieli/ai-agent-book · ★ 48k ▲2.3k/7d · Python

AI Agent Book is a Chinese-first, open-source textbook and experiment collection about context, tools, memory, evaluation, model training, and multi-agent systems. An English community translation exists in web, PDF, EPUB, and source form, but the README warns that translations can lag behind the Chinese original.

Verdict Our AI Agent Book environment occupied 6,774 MB, the build failed in 10 seconds, and tests ended with 41 collection or setup errors, so this is a reading library with independent labs rather than one reproducible application. Use it chapter by chapter, especially if you can read the Chinese source and want wide coverage of agent engineering.

openhuman

tinyhumansai/openhuman · ★ 40k ▲227/7d · Rust

OpenHuman is an early-beta desktop assistant that keeps a local memory of your files and connected accounts, then uses agents to research, automate work, and answer through chat or messaging channels. It combines an Obsidian-readable memory store, visual workflows, model routing, and a broad tool catalog in one application.

Verdict Our build finished in 55 seconds after installing 1,313 packages and 897 MB, but the test run still had no final result at the 900-second cap, so OpenHuman is an ambitious beta rather than a dependable personal data hub. Try the installer if you specifically want local memory plus visual agent workflows and can use a separate profile with low-value accounts.

code-review-graph

tirth8205/code-review-graph · ★ 31k ▲200/7d · Python

code-review-graph builds a local SQLite graph of functions, classes, imports, calls, tests, and changed files, then exposes focused slices through a CLI and MCP server. Its purpose is to stop coding agents from rereading an entire repository when they only need the code and tests affected by a change.

Verdict Our install consumed 5,344 MB and the test suite reached only 16% before a 900-second timeout, so code-review-graph is a deliberate toolchain addition rather than a lightweight MCP extra. Trial it on a large repository when agent context selection is a measured problem, and compare every proposed impact path with code review evidence.

FastGPT

labring/FastGPT · ★ 30k ▲41/7d · TypeScript

FastGPT's primary README and latest release notes are written in Chinese, though an English README is available. FastGPT is a visual platform for building AI agents, retrieval-backed chat applications, knowledge bases, and tool workflows, with cloud and self-hosted ways to run them.

Verdict Our FastGPT install pulled 2,548 packages and used 2,445 MB, then its test command failed because `@fastgpt/admin` was missing from the workspace filter. Trial it if you need a Chinese-first, self-hosted agent and knowledge platform and have operators for its storage, migrations, model keys, and sandbox.

GenAI_Agents

NirDiamant/GenAI_Agents · ★ 24k ▲89/7d · Jupyter Notebook

GenAI Agents is a collection of more than 50 Jupyter notebook tutorials that show how different language-model agents are assembled. It is a learning library for comparing patterns such as LangGraph workflows, multi-agent research, RAG, tool use, and MCP, rather than a single agent you deploy.

Verdict Our GenAI Agents install took 69 seconds, pulled 119 packages, and returned 137 known vulnerability findings, so it is useful as a reading lab but a poor production dependency. Use it to compare 50-plus agent ideas and borrow small, reviewed pieces.

Skill_Seekers

yusufkaraaslan/Skill_Seekers · ★ 15k ▲40/7d · Python

Skill Seekers is a Python tool that turns websites, repositories, office files, videos, and chat exports into reference packs for AI assistants and retrieval systems. It handles the scraping, organization, optional AI editing, and packaging that teams otherwise wire together for each source and destination.

Verdict Our Skill Seekers run installed 148 packages and passed 4,058 tests, but it also produced 14 failures, 16 collection or setup errors, and 37 known vulnerabilities. That makes it an interesting choice for teams that truly need its 18 input types and 22 outputs, provided they can audit dependencies, inspect generated references, and keep MCP access private.

fastapi_mcp

tadata-org/fastapi_mcp · ★ 12k ▲6/7d · Python

FastAPI-MCP turns selected routes from an existing FastAPI application into tools that an AI client can call through Model Context Protocol. It keeps request schemas, route documentation, and FastAPI dependencies in the path, which saves teams from maintaining a second tool server by hand.

Verdict Our FastAPI-MCP run finished with 12 of 66 tests passing, while 32 failed and 22 stopped during collection or setup, so a production trial needs its own end-to-end transport check. The idea is attractive for a FastAPI team with carefully selected routes, pinned dependencies, and explicit authentication.

MemOS

MemTensor/MemOS · ★ 11k · TypeScript

MemOS is an English-documented memory layer for AI agents, with a Chinese README translation. It stores, searches, edits, and shares long-term memories through a hosted API, a self-hosted service, or local plugins for OpenClaw, Hermes Agent, and DeepSeek Harness.

Verdict Our MemOS run passed 936 tests, but 6 failed and pip-audit found 29 known vulnerabilities after a 6,217 MB install, so the current checkout misses a strict production gate. The local plugin is the most approachable path for an existing OpenClaw, Hermes, or DeepSeek Harness user.

openstatus

openstatusHQ/openstatus · ★ 9.1k ▲24/7d · TypeScript

OpenStatus combines uptime checks and public status pages in one open-source system. It can run as a managed service or on your infrastructure, with an API, CLI, Terraform provider, monitoring-as-code files, and an MCP server sharing the same workspace credentials.

Verdict Our OpenStatus install pulled 1,538 packages and occupied 1,832 MB, then both the 6-second build and 7-second test run failed, so source contributors should expect setup investigation. The product is appealing for teams that want monitors, status pages, Terraform, CLI, and MCP under one API.

lamda

firerpa/lamda · ★ 8.3k ▲17/7d · Python

FIRERPA is an Android control and automation system with an on-device service and a Python client. It combines remote desktop, UI automation, traffic inspection, Frida tooling, fleet networking, and an MCP endpoint for agents; English and Chinese documentation are linked from the README.

Verdict Our FIRERPA client run installed 44 packages in 17 seconds and built in 9 seconds with 0 known audited vulnerabilities, but the repository exposed no test target for us to run. Use it when one dedicated Android control plane can replace several separately operated tools and your team can secure the device service.

AI-Infra-Guard

Tencent/AI-Infra-Guard · ★ 6.4k ▲188/7d · Python

AI-Infra-Guard is an English-documented platform for testing AI services, agents, Model Context Protocol servers, agent skills, and language models; its README links eight translations. It puts live-service vulnerability checks, source inspection, jailbreak exercises, and model API checks in one self-hosted web app.

Verdict Our AI-Infra-Guard run installed 301 packages and built successfully, but 7 of 16 tested packages failed, so it belongs in an evaluation lab before it belongs in a security process. Use it when one isolated team needs its six scanning and evaluation jobs and can verify every finding.

maka

apache/maka · ★ 5.5k ▲344/7d · TypeScript

Apache Maka (Incubating) is a local-first coding-agent workspace with desktop, terminal, and evaluation interfaces. It keeps model messages, tool calls, permissions, results, and stop events in a durable local record so a run can be inspected or recovered instead of living only in a chat window.

Verdict Our Maka run installed 849 packages and built successfully, but its test command exited 1 after 801 seconds with 3 of 1,095 node:test cases failed. The durable event log and shared Runtime Host are worth studying for agent work that must survive refreshes and crashes.

dsh-TUI

ccch1mneyyy/dsh-TUI · ★ 3.1k ▲142/7d · TypeScript

dsh-TUI is a Chinese-first terminal interface plugin for DeepSeek Harness, with a full English README and English reference pages. It gives the official DSH agent a Claude Code-style conversation screen, tool cards, session controls, context meters, rewind, themes, MCP status, and a VS Code companion without patching the DSH core.

Verdict Our dsh-TUI install completed in 33 seconds, but its source build failed after 41 seconds on an exact pnpm 11.21.0 versus 11.23.0 check, and the repository exposed no test target. Existing DeepSeek Harness users may still find it the best interface upgrade because it preserves DSH services while adding serious session navigation and visibility.

ADR

uber/ADR · ★ 1.6k ▲22/7d · Python

ADR means Agentic AI Detection and Response, an Uber project for inventorying AI tools, collecting agent activity, benchmarking defenses, and detecting suspicious sessions. It covers employee tools such as Claude Code, Cursor, and Codex as well as customer-facing agents, but the open release does not include prevention.

Verdict Our ADR Detection environment occupied 6,219 MB, pip-audit reported 100 known vulnerabilities, and test collection stopped with 61 tests passed and 1 marker error. Use the Sensor and benchmark as security engineering material when you can isolate the research environment and govern sensitive agent telemetry.

Konnect

mixelpixx/Konnect · ★ 684 ▲127/7d · Rust

Konnect is a native KiCAD 10 plugin that lets Claude and other AI assistants create and inspect circuit schematics and PCB layouts through the Model Context Protocol. It replaces a multi-language predecessor with one Rust binary, while covering checks, routing support, part search, and manufacturing exports.

Verdict Our source build completed in 273 seconds, and all 3,042 tests passed. Konnect is unusually ambitious and technically convincing for a beta, especially if KiCAD 10 is already your board-design standard.

lobehub

lobehub/lobehub · ★ 83k ▲173/7d · TypeScript

LobeHub is an agent workspace for creating AI assistants, organizing them into groups and projects, scheduling work, and connecting them to models, messaging channels, files, and MCP-compatible tools. Its primary README is in English, and a Simplified Chinese README is also available.

Verdict Our LobeHub install spent 296 seconds before npm rejected `workspace:*`, so the first decision is whether your team will follow its pnpm workflow and accept a 198.4 MB monorepo. Trial it if agent groups, scheduled work, messaging, memory, and MCP tools genuinely belong in one workspace.

voicebox

jamiepine/voicebox · ★ 54k ▲1.1k/7d · TypeScript

Voicebox is a local AI voice studio that combines voice cloning, speech generation, dictation, transcription, audio editing, and agent speech in one desktop app. It runs several speech engines on your machine and exposes REST and MCP interfaces so local applications and coding agents can use the same voices and captures.

Verdict Our Voicebox install pulled 408 packages and used 911 MB, then the build failed because `python` was unavailable and there was no test target to run. The app is worth a careful trial for local voice I/O, especially if REST or MCP access matters, but it asks the operator to own models, hardware, and output validation.

ai-engineering-hub

patchy631/ai-engineering-hub · ★ 38k ▲111/7d · Jupyter Notebook

AI Engineering Hub is a collection of more than 90 small projects and tutorials covering RAG, agents, model comparisons, multimodal applications, fine-tuning, and MCP. It is a learning library rather than one deployable product; our lab measured its `Multi-Agent-deep-researcher-mcp-windows-linux` example, which combines 3 CrewAI agents, Linkup web search, local DeepSeek R1 through Ollama, Streamlit, and an MCP server.

Verdict Our measured researcher installed 33 packages in 30 seconds and built in 11 seconds, but it exposed no test target and the repository has 0 CI workflows. Use AI Engineering Hub as a menu of code samples, especially when you want to compare several approaches quickly.

AiToEarn

yikart/AiToEarn · ★ 26k ▲154/7d · TypeScript

AiToEarn's primary README and latest release notes are in Chinese, and a complete English README is provided. It is a content operations system for generating media, scheduling and publishing it across social networks, automating engagement, and connecting those actions to agents through MCP.

Verdict Our AiToEarn backend install pulled 1,487 packages and 986 MB in 58 seconds, then offered no build or test target to verify. Use it only if its unusual combination of Chinese and international channels, content generation, and MCP control saves enough work to justify the relay, credential, and platform-risk review.

eigent

eigent-ai/eigent · ★ 15k ▲58/7d · TypeScript

Eigent is an Electron desktop application for assigning work to one AI agent or a coordinated group of agents. It gives those agents a browser, terminal, local files, scheduled workflows, skills, and MCP connectors, with either cloud-backed or locally deployed services.

Verdict Our Eigent run used 1,634 MB, found 27 known vulnerabilities, and ended with 34 of 739 tests failing, so it is an evaluation build rather than a safe default for unattended desktop work. Try it if you specifically want CAMEL-based multi-agent work in an Electron interface and can confine its browser, terminal, files, and credentials.

corsair

corsairdev/corsair · ★ 11k ▲7/7d · TypeScript

Corsair is a TypeScript integration layer that lets agents call third-party services without receiving the underlying API keys. It wraps service plugins with OAuth or key storage, tenant scoping, webhook handling, cached data, and per-endpoint policies that can allow, deny, or pause an action for human approval.

Verdict Our Corsair checkout installed 2,105 packages and consumed 2,278 MB, then 34 of 43 Jest tests failed even though the 565-second build passed. Its permission and credential model addresses a real agent risk, but the measured suite result is too weak for blind production adoption.

computer

cloudflare/computer · ★ 9.2k ▲53/7d · TypeScript

Cloudflare Computer gives an agent a persistent filesystem inside a Durable Object and lets it run commands or JavaScript against those files. The same workspace can use a Worker shell, an isolated JavaScript runtime, or a full Linux container, but Cloudflare labels the package preview-only and unsuitable for production.