mrkeyoor.comAsk

volcengine/OpenVikingtested 13 Oct 2026commit 92285e5

OpenViking review

Pilot OpenViking when inspectable context justifies operating a separate service.

  • Build 6 s
  • Tests failed
60 sto install
  • 35 packages
LLM Toolsevaluationupdated 20 Oct 2026
+255stars / 7d

OpenViking is a context database for AI agents that stores documents, memories, and skills in a browsable virtual filesystem. Its viking:// paths let developers inspect and edit retained context. English documentation is available.

Verdict

Pilot OpenViking when inspectable context justifies operating a separate service. commit 92285e5 installed and built in our sandbox, but tests stopped on a missing pytest_asyncio import before exercising retrieval. Require a successful search of a known document and an inspection of captured content before connecting everyday sessions.

We ran it

Lab card: what happened when we ran OpenViking
Lab card from our sandbox run
Screenshot of OpenViking (openviking.ai)
openviking.ai
Install✓ · 60s35 packages · 38 MB
Build✓ · 6s
Tests✗ · 2sran, no count parsed
Known vulns0(pip-audit)
Repo5069 files~1,115,732 lines of source · 103.5 MB · 30 CI workflows · Dockerfile · tests dir

Answers from our run

Does OpenViking build from source?

Dependencies installed in 60 seconds (35 packages), and the build succeeded in 6 seconds. We cloned commit 92285e5 into a clean Debian container with 3 CPUs and no project-specific setup.

Do OpenViking's tests pass?

No. The project's own test command exited unsuccessfully after 2 seconds, and its output did not expose an individual test count.

Does OpenViking have known vulnerabilities in its dependencies?

pip-audit found none in the dependency tree at the time of our run.

Who should not use OpenViking?

Applications that require a strict total search-result limit without handling duplicates: the retrieval guide says merged results may exceed the per-query limit and need not be deduplicated.

What are the alternatives to OpenViking?

Mem0, LangMem. Pilot OpenViking when inspectable context justifies operating a separate service.

Setup4/5Our source checkout installed in 60 s and built in 6 s with 35 dependencies, but we did not run a README task.Editor’s note: The recorded install and build succeeded, but model configuration and a usable server were not exercised, so setup ease remains unassessed.
Docs4/5The README has an install section (“Quick start”) and usage in the same section; we did not run its quickstart ourselves.Editor’s note: The supplied configuration and retrieval pages explain important boundaries, but their instructions and integration paths were not validated sufficiently to support a numerical rating.
Community5/5Last push on 2026-10-20 and 42,236 stars.Editor’s note: Supplied metadata records a push at 2026-10-20T04:45:12Z, observed October 20, 2026. A recent push and one issue discussion do not establish overall maintenance health.
Maturity2/5Latest release v0.5.1 on 2026-10-19 (pre-1.0), repository created 2026-01-05 and its test suite did not pass in our sandbox.Editor’s note: The suite stopped on an import before exercising the product. Retrieval quality, multi-user operation, and recovery were not assessed in the supplied run.

How we score

Who it’s for

Agent developers who want stored context they can browse and edit through named paths.
Teams prepared to run a memory server and configure its models and storage.
Developers comparing direct retrieval with session-aware search on their own documents.

Who it’s NOT for

Applications that require a strict total search-result limit without handling duplicates: the retrieval guide says merged results may exceed the per-query limit and need not be deduplicated.
Teams expecting the memory-plugin installer to run natively on Windows: the README directs Windows users to its beta desktop app.
Claude Code users who need verified conversation-only capture before a pilot: issue 6314 reports duplicated turns containing system-prompt context with OpenViking 0.5.1, plugin 0.7.12, and Claude Code 2.1.295 or later.
Teams requiring permissive licensing for the main server project: the README identifies its license as AGPLv3 and lists separate licenses for some components.

Setup reality

Our sandbox installed OpenViking commit 92285e5 in 60 seconds, recording 35 dependencies and 38 MB on disk. The build succeeded in 6 seconds. Tests exited 4 after 2 seconds while loading tests/conftest.py because pytest_asyncio was missing. Pip-audit reported 0 known vulnerabilities; that result does not establish application security.

The README's self-hosted setup requires an embedding model and a vision-language model. The setup wizard writes ~/.openviking/ov.conf and prompts for those settings separately. Hosted API services need suitable credentials; local Ollama is also documented. The current configuration guide additionally describes a default local embedding model that requires the local inference dependency. We did not exercise these model configurations.

The recorded 2026-10-13 sandbox configuration used 3 CPUs, 8 GB RAM, and image ghcr.io/astral-sh/uv:python3.12-bookworm. Actual runtime versions require the raw toolchain logs. The current memory-plugin installer requires macOS or Linux, Node.js 18+, and curl. Windows users are directed to the beta desktop app.

Stored memories have paths you can inspect

OpenViking gives retained context a viking:// address. Its README describes a virtual filesystem where developers can inspect and edit documents, memories, and skills. Consider it when seeing what an agent retained would help you operate that agent. Our sandbox run established installation and build results, but did not establish working retrieval.

The README separates directory abstracts, longer overviews, and full source content into loading levels. An agent can read a summary before opening the underlying file. That gives operators material to inspect, while leaving them responsible for checking whether generated summaries and retained memories represent their sources accurately.

What happened when we ran it

Our sandbox installed OpenViking commit 92285e5 in 60 seconds, recorded 35 dependencies using 38 MB on disk, and completed the build in 6 seconds. Tests stopped after 2 seconds with a missing pytest_asyncio import. The command exited 4 while loading tests/conftest.py, before the suite could exercise retrieval.

The supplied traceback identifies line 11 of conftest.py and ends with ModuleNotFoundError for pytest_asyncio. It does not show why the module was missing, whether a documented development configuration would supply it, or whether application tests would pass after it became available.

The 2026-10-13 run recorded 3 CPUs, 8 GB RAM, and image ghcr.io/astral-sh/uv:python3.12-bookworm. The image label does not verify actual runtime versions. The checkout had 5,069 files, about 1,115,732 source lines, and occupied 103.5 MB. The run recorded 30 CI workflow files, a Dockerfile, Compose configuration, and a tests directory. Pip-audit reported 0 known vulnerabilities. That audit does not measure retrieval quality or establish service security.

These measurements concern commit 92285e5. The current documentation and the later plugin report below are separate evidence, not behavior verified on that checkout.

Models, access, and licensing affect adoption

The configuration guide uses ~/.openviking/ov.conf and configures embeddings separately from the vision-language model. API-based providers need appropriate credentials. The README also documents local Ollama. The configuration guide describes a default local embedding model, with a separate inference dependency and a model download on first use unless cached. These are documented setup options; our lab did not verify model calls or document processing.

Changing ov.conf requires a restart. The guide distinguishes startup configuration from a limited set of runtime settings. Choose the models and persistent workspace before treating a package build as a usable memory system.

For authenticated agent integrations, the README requires a user key rather than a root key for reading and writing memories. It also directs operators to configure authentication before exposing the server beyond localhost. Check the setup path for your intended client, including the installer's platform requirements.

The README identifies the main project as AGPLv3. It lists Apache 2.0 for crates/ov_cli and examples, while the Hermes plugin retains MIT licensing and framework adapters declare AGPLv3. Check the component you intend to use or modify against your licensing requirements.

Search can return duplicates beyond the requested limit

The retrieval guide describes global vector retrieval with optional reranking. Direct find requests do not use session context. Search can use session context and an LLM for intent analysis when enabled. Target directories constrain search scope, but retrieval does not navigate directories recursively.

For search, the limit applies to each planned query. Merged results may exceed that limit, and deduplication across queries is not guaranteed. An application with a fixed context allowance should inspect the returned set before passing it to a model.

Start a pilot with a document whose answer you know and another document that should be irrelevant. Check the returned source paths, then compare direct retrieval with a session-aware request. Inspect result counts and duplicates as well as relevance. These are proposed acceptance checks, not experiments completed in our sandbox.

Inspect Claude Code captures before routine use

Issue 6314, opened October 19, 2026, reports that the Claude Code memory plugin turns API-request records into conversation turns. The reporter lists OpenViking 0.5.1, plugin 0.7.12, Windows, and Claude Code 2.1.295 or later. The reported duplicates contain system prompts and injected context. This plugin report is separate from our Python test-startup failure.

An October 20 comment points to a proposed shared-parser fix and reports validation of a synthetic request-envelope case. The commenter explicitly says this does not verify the reporter's Windows transcript or a live capture. The comment also says existing capture cursors and previously archived duplicates are not migrated. Inspect a small captured archive before connecting routine sessions, and check that every retained field belongs there.

The supplied repository metadata records a push at 2026-10-20T04:45:12Z, observed on October 20, 2026. That dated activity and the issue discussion do not establish overall maintenance health or resolution of this capture report.

Compare deployment and storage choices

LangMem provides functional memory primitives and native integration with LangGraph's storage layer. Consider it for an existing LangGraph application when using that storage integration is the priority.

Mem0 presents library, self-hosted server, and cloud options. Compare those deployment paths with the service you would operate for OpenViking. No comparative retrieval measurements are supplied for either alternative.

OpenViking merits a narrow pilot when browsing retained context would help your team operate an agent. Before connecting a daily workflow, retrieve a known document and inspect a captured session archive. The supplied build result supports starting that evaluation; the blocked tests leave application behavior unresolved.

Alternatives

ProjectWhat it isPick it when
Mem0 gh↗An agent memory project with library, self-hosted server, and cloud options.Compare this when choosing among library, self-hosted server, and cloud deployment is your main decision.
LangMemMemory primitives with native integration with LangGraph's storage layer.Consider this when adding memory to an existing LangGraph application and using its storage integration is the priority.

Sources

  1. OpenViking README
  2. OpenViking retrieval mechanism
  3. OpenViking model and service configuration
  4. Claude Code capture report
  5. Mem0 README
  6. LangMem README

More llm tools reviews