mrkeyoor.com_
Tue 01 Sept 17:40 UTC
Automationevaluationupdated 26 Aug 2026

sim review

Sim is a shared workspace for building, deploying, and monitoring AI agents and automated workflows. Teams can assemble work visually, through chat, or in code, then connect models, business apps, files, tables, and searchable knowledge in one system.

+25stars / 7d
Verdict

Our Sim checkout installed 2,519 packages and used 2,560 MB, then both the 8-second build and 203-second test command exited 1. Trial the published self-hosting images if one workspace for agents, data, schedules, and run monitoring would replace several internal tools. Avoid adopting the source tree casually: its desktop prebuild, wide monorepo test surface, credentials, and supporting services demand a real platform owner.

We ran it

Lab card: what happened when we ran simScreenshot of sim (www.sim.ai)
Install✓ · 117s2519 packages · 2560 MB
Build✗ · 8s
Tests✗ · 203s116 passed · 0 failed of 116 (vitest)
Repo18392 files~3,044,743 lines of source · 294.9 MB · 15 CI workflows

Answers from our run

Does sim build from source?

Dependencies installed in 117 seconds (2519 packages), and the build failed. We cloned commit f1a3456 into a clean Debian container with 3 CPUs and no project-specific setup.

Do sim's tests pass?

Yes: 116 of 116 passed when we ran the project's own test command (vitest). Some failures need services or credentials a bare container does not have.

Who should not use sim?

Small self-hosted boxes: the setup guidance warns that the stack needs at least 12 GB of RAM.

What are the alternatives to sim?

n8n, Langflow, Dify. Our Sim checkout installed 2,519 packages and used 2,560 MB, then both the 8-second build and 203-second test command exited 1.

Setup2/5Wizard helps, but source build and tests failed in our run
Docs4/5Clear wizard, service, secret, model, and management commands
Community5/529,471 stars with an August 26 push and active release work
Maturity3/5Broad v0.8.11 platform with failing checks in our sandbox

Discussed on

  1. hnShow HN: Sim – Apache-2.0 n8n alternative240 points
  2. hnShow HN: Sim Studio – Open-Source Agent Workflow GUI196 points

Who it’s for

Product and operations teams building multi-step AI workflows with developer oversight.
Organizations that want a visual builder, shared data, execution logs, schedules, and deployed endpoints in one workspace.
Self-hosters comfortable operating Docker, PostgreSQL, secrets, model credentials, and external integrations.
Developers who want to extend a TypeScript monorepo built around Next.js, Bun, Drizzle, and realtime services.

Who it’s NOT for

Small self-hosted boxes: the setup guidance warns that the stack needs at least 12 GB of RAM.
Fully offline installations that need the Chat surface: the README says Chat is Sim-managed and connects through a Chat API key.
Teams expecting every integration to work without credentials: model providers and business services still need their own keys or OAuth setup.
Contributors who require a clean build from a generic Node image: our desktop build stopped in a prebuild-fetch script after 8 seconds.
Release gates that cannot accept a failing test command: our run exited 1 after 203 seconds, with app failures and a desktop timeout in the log tail.
Buyers wanting a small codebase: the checkout held 18,392 files, about 3,044,743 source lines, and installed 2,519 packages.

Setup reality

Our sandbox installed 2,519 packages in 117 seconds and used 2,560 MB on disk. The build failed in 8 seconds inside the desktop package's ensure-pty-prebuilds.ts fetch path. Four of 7 build tasks succeeded before the command exited 1.

The test command also exited 1 after 203 seconds. The harness summary recorded 116 Vitest tests passed and 0 failed, while the monorepo log tail separately showed 3 app failures and 1 desktop timeout. The repository had no top-level tests directory.

The documented self-host route needs Node.js 20+, Docker, generated secrets, PostgreSQL, and a Chat API key if Chat is used. Optional email, storage, sandbox, jobs, cache, knowledge, LLM, and OAuth integrations add their own services and credentials.

One workspace holds agents, data, and run history

Sim combines a visual workflow builder with chat, code blocks, integrations, files, tables, and searchable knowledge. A team can connect model calls to business systems, deploy the workflow, then inspect its runs and logs without building a separate control panel. The current README claims more than 1,000 integrations, but each useful provider still needs credentials and an acceptance test against the actions the workflow will take.

The scale is closer to an application suite than a canvas component. Our commit f1a3456 checkout contained 18,392 files and about 3,044,743 lines of source in 294.9 MB. The monorepo includes the main web app, desktop work, realtime services, shared packages, setup tooling, and documentation. Next.js, Bun, PostgreSQL, Drizzle, Socket.io, and Trigger.dev are among the named pieces.

The setup wizard creates a multi-service installation

For normal self-hosting, the README recommends npx sim-setup rather than cloning the repository. The wizard creates a deployment directory, generates secrets, provisions the database, writes environment settings, connects a Chat API key, and starts published images with Docker Compose. It also has commands for status, logs, diagnosis, updates, restart, and removing containers while retaining data. That is a useful operational surface for a large stack.

Requirements start with Node.js 20 and Docker. The README's setup link warns that insufficient memory means less than 12 GB of RAM. PostgreSQL is part of the documented stack, and the setup tool can add email, object storage, sandboxing, jobs, cache, knowledge, LLM, or an integration such as Slack. Each addition changes the secrets, network access, backup plan, and failure modes an operator owns.

What happened when we ran it

Our unprivileged Node 22 sandbox installed the repository in 117 seconds. Bun added 2,519 packages, leaving 2,560 MB on disk. The build failed after 8 seconds. Its final lines point to apps/desktop/scripts/ensure-pty-prebuilds.ts, where fetchPrebuild raised the error; the desktop build then exited 1. Four of 7 tasks had succeeded at that point.

The log tail does not say why the prebuild fetch failed, so blaming network policy, a missing artifact, or application code would be guesswork. The repository scan found 15 CI workflow files, no Dockerfile at the scanned location, no top-level tests directory, and monorepo workspaces. The 8-second result applies to commit f1a3456 in our stated container, not the published Compose images.

Our test command also exited 1 after 203 seconds. The harness summary says Vitest ran 116 tests with 116 passed and 0 failed, while the wider monorepo log tail reports 3 failures in an app compiler file and 1 timed-out desktop browser-import test. The desktop package's summary showed 1 failed file and 60 passed files. These outputs describe different slices of the same command, so the defensible overall result is failure.

The desktop timeout was set to 10,000 ms and occurred in chromium-site-names.test.ts. The supplied log does not establish whether the test is flaky, the sandbox is slow, or the code has a defect. The 3 app failures also need their full assertion messages before diagnosis. A maintainer evaluating this commit should rerun the exact failing package tests and preserve the complete logs.

Chat remains a Sim-managed service when self-hosted

Self-hosting does not make every feature local. The README says Chat is managed by Sim and requires a Chat API key connected during setup. Keys can later be viewed, created, or revoked on sim.ai. Local models are supported through Ollama and vLLM, but the presence of those inference routes does not remove the Chat account boundary described by the project. Air-gapped buyers should map each required feature before installing.

The setup tool reports configured, missing, or invalid capabilities without printing credential values. That distinction is useful once an installation has several optional services. A 2,519-package codebase can fail far from the workflow canvas, so health checks should cover PostgreSQL, realtime connections, scheduled work, storage, model endpoints, and each OAuth callback rather than only loading port 3000.

Release v0.8.11 is active and still pre-1.0

Release v0.8.11 was published on August 25, 2026, and the repository was pushed again on August 26. GitHub showed 29,471 stars and 289 combined issues and pull requests when fetched. The release included changes to setup, secrets, tables, OAuth credentials, webhooks, vLLM endpoints, file handling, and CI audits. That breadth shows current maintenance and also a wide regression surface.

Apache 2.0 covers the repository according to the current README and GitHub license result. The product is still on a 0.x line, so pin a release and rehearse upgrades against representative workflows. Our 117-second install succeeded, but the failed build and test command prevent a clean source recommendation for f1a3456. Published images are the sensible evaluation route for users; contributors need to resolve or reproduce both failures first.

Sim earns its weight only when it replaces several tools

A team using agents, schedules, integrations, shared files, tables, knowledge, and monitoring can justify Sim's 2,560 MB dependency environment and service stack. The setup wizard reduces initial friction, while the management commands give operators a coherent way to inspect the installation. Those benefits matter more as workflow count and contributor count grow.

For a handful of model chains, this is too much machinery. The source tree failed both major gates in our sandbox, Chat keeps a hosted key boundary, and optional capabilities multiply credentials. Pick Sim when someone will own it as an internal platform. Otherwise, n8n, Langflow, Dify, or a small coded service may be easier to audit and keep running.

Alternatives

ProjectWhat it isPick it when
n8n gh↗A visual automation platform with a large connector catalog and AI workflow support.pick this instead when conventional business automation matters more than an agent-centered workspace.
Langflow gh↗A Python visual builder for language-model and retrieval flows.pick this instead when the team is Python-first and wants a narrower model workflow layer.
Dify gh↗A platform for building agent workflows, retrieval applications, and deployed AI endpoints.pick this instead when packaged AI application management and retrieval are the main requirements.

What people are saying

  1. [github-trending] simstudioai/sim
  2. [hackernews] Show HN: Silent Shark – tactical map-based WWII submarine sim
  3. [theverge] Tomodachi Life: Living the Dream is a quirky life sim that’s worth buying at this discount

Sources

  1. Sim README
  2. Sim v0.8.11 release
  3. Sim self-hosting documentation
  4. Sim environment variables reference

More automation reviews

rclone · lego · OpenCLI · web-access · Karabiner-Elements · WiiUDownloader · the whole board →