Devtoolwatch BoardGet it every Monday

OpenViking review

OpenViking is an open-source context database that stores AI agent knowledge, memory, and skills in one virtual filesystem under viking://.

Updated Nov 1, 2026. Review written Oct 22, 2026 from the project's own sources; usage data refreshed every Monday.

TL;DR

OpenViking is an open-source context database for AI agents that organizes resources, memories, and skills as a browsable virtual filesystem. It is for developers who want agent memory they can inspect and edit. It differs by scoping semantic search to directories and generating layered summaries so agents read abstracts before full content.

OpenViking at a glance

Category
RAG, data and memory
Maintained by
Volcengine (organization, @volcengine on GitHub)
License
AGPL 3.0
Language
Python, TypeScript, Rust
Repository
github.com/volcengine/OpenViking
Website
openviking.ai
PyPI package
openviking 0.5.1.1
Latest release
v0.5.1.1 on GitHub, Oct 20, 2026
First published
Jan 5, 2026

What OpenViking does

  • OpenViking organizes agent context as a virtual filesystem that agents navigate with ls, tree, read, write, and grep.
  • OpenViking scopes semantic search to a project or memory subtree instead of scanning a flat vector pool.
  • OpenViking generates L0 abstracts and L1 overviews per directory so agents judge relevance before loading full L2 content.
  • OpenViking archives committed sessions and extracts memories as Markdown files you can inspect, edit, and merge.
  • OpenViking provides a memory plugin installer for Claude Code, Codex, Cursor, TRAE, OpenCode and other coding agents.
  • OpenViking offers Python, Go, and TypeScript SDKs plus an HTTP API for custom integrations.
  • OpenViking with VikingBot enabled can compile source material into a wiki, knowledge graph, or report via ov compile.

How it works

OpenViking runs as a server, either self-hosted or as the Volcengine-hosted OpenViking Service. The server needs a model provider with an embedding model and a VLM. Agents connect through hooks, MCP, plugins, or SDKs, and the ov CLI imports resources and runs find, search, and grep against viking:// URIs.

How to install and start using OpenViking

Server with uv

uv tool install openviking --upgrade && openviking-server init

Agent memory plugin

curl -fsSL https://openviking.ai/install | bash

With VikingBot

pip install "openviking[bot]"

Requirements

  • Self-hosting needs uv and Python 3.10 or newer.
  • A model provider with an embedding model and a VLM, such as Volcengine, OpenAI, Kimi, GLM, or local Ollama.
  • The plugin installer needs macOS or Linux, Node.js 18+, and curl; Windows users use the desktop app.
  • If server auth is on, agents need a user key; a root key cannot read or write memories.

First steps

  1. Run openviking-server init to write ~/.openviking/ov.conf and configure your model provider; keep the foreground server terminal open.
  2. Import a repository with ov add-resource, then check progress with ov task status until it completes.
  3. Browse and query imported context with ov ls, ov tree, and ov find.
  4. Install the memory plugin, restart your agent, and ask it to remember a preference you check in a later session.

Who OpenViking is for

  • Developers who want cross-session memory for coding agents like Claude Code, Codex, or Cursor.
  • Teams who need to inspect and edit what their agent has stored.
  • Agent builders combining documents, code, memories, and skills in one retrieval store.
  • Organizations that want to self-host with accounts, user isolation, and resource ACLs.

Limitations to know

  • OpenViking is licensed AGPL-3.0, which affects how you can embed or redistribute it.
  • OpenViking processes memories in the background, so a new memory may not be retrievable right away.
  • OpenViking's distributed deployment and official support require the licensed Self-Managed edition.
  • OpenViking v0.5.1 includes API removals, write-semantics and storage-format changes; read migration notes before upgrading.

Is OpenViking growing? Usage data

In the week of Oct 25–31, 2026, OpenViking ranked #235 on the Devtoolwatch Board and #25 in RAG, data and memory, with a momentum score of 39 out of 100. Its main package had 66.8k downloads that week, −35% against its own 4-week average.

New GitHub stars per week for OpenViking12 weeks from Aug 9, 2026: 425, 3.6k, 2.2k, 1.4k, 1.2k, 1.3k, 653, 474, 486, 628, 599, 541 new stars, an average of 1.1k a week.Aug 9week of Oct 25
New GitHub stars per week for the last 12 weeks, from GitHub's stargazer history (today's stargazers by the week they starred).
Show the weekly numbers
New GitHub stars per week for OpenViking
Week ofNew stars
Aug 9, 2026425
Aug 16, 20263.6k
Aug 23, 20262.2k
Aug 30, 20261.4k
Sep 6, 20261.2k
Sep 13, 20261.3k
Sep 20, 2026653
Sep 27, 2026474
Oct 4, 2026486
Oct 11, 2026628
Oct 18, 2026599
Oct 25, 2026541

The Board ranks tools by four signals each week: new stars, new stars for the repo's size, added downloads and download growth beyond the market. How the ranking works.

OpenViking alternatives

Other RAG, data and memory tools on the Board in the week of Oct 25–31, 2026, by momentum:

RAG, data and memory ranked by momentum, the week of Oct 25–31, 2026
RankToolStarsDownloads/wk
#1XERJXERJ is a local-first search engine and MCP server for AI agents, packaged as one Rust binary with an Elasticsearch-compatible API.17.9k–
#2MemvidMemvid is a single-file memory layer for AI agents that packages data, embeddings, search indexes, and metadata into one portable .mv2 file.17.3k–
#3LEANNLEANN is a Python vector database and RAG library that recomputes embeddings on demand to keep indexes small on personal devices.13.7k–
#4GraphifyGraphify is an AI coding assistant skill that turns a codebase, with its docs, PDFs, images and videos, into a queryable knowledge graph.140k480.7k
#5MarkItDownMarkItDown is a lightweight Python utility from Microsoft that converts PDFs, Office documents, images, audio and other files to…193.2k4.7M
#6FirecrawlFirecrawl is an API and open source project to search, scrape, crawl, and interact with web pages, returning LLM-ready markdown or JSON…197.4k2.9M

All RAG, data and memory tools, ranked

Recent OpenViking releases

OpenViking FAQ

Is OpenViking free and open source?

Yes. OpenViking's source code is on GitHub (volcengine/OpenViking) under the AGPL 3.0 license, so you can use it for free.

How do I install OpenViking?

Run uv tool install openviking --upgrade && openviking-server init. The openviking package needs Python 3.10 or newer. The install section above has the other options from the project's README.

Is OpenViking still maintained?

Yes. The volcengine/OpenViking repository was last updated on Oct 21, 2026, and the latest release, v0.5.1.1 on GitHub, came out on Oct 20, 2026.

How popular is OpenViking?

OpenViking has 41.2k GitHub stars as of Nov 1, 2026. It gained 541 new stars in the week of Oct 25–31, 2026, against a 12-week average of 1.1k a week. Its main package had 66.8k downloads that week, −35% against its own 4-week average. That made it #25 among RAG, data and memory tools on the Devtoolwatch Board.

OpenViking vs Mem0: how do they compare?

OpenViking and Mem0 are both tracked by Devtoolwatch as RAG, data and memory tools. OpenViking has 41.2k GitHub stars and Mem0 has 68.2k. In the week of Oct 25–31, 2026, OpenViking's main package had 66.8k downloads and Mem0's had 884.1k. Mem0 is a memory layer for AI agents and apps that stores and retrieves user, session, and agent context across interactions. Mem0 review.

OpenViking vs Honcho: how do they compare?

OpenViking and Honcho are both tracked by Devtoolwatch. OpenViking has 41.2k GitHub stars and Honcho has 7.9k. Honcho review.

What are the best alternatives to OpenViking?

Other RAG, data and memory tools on the Devtoolwatch Board in the week of Oct 25–31, 2026, by momentum: XERJ (#1), Memvid (#2), LEANN (#3), Graphify (#4), MarkItDown (#5). The alternatives table above has their numbers.

How is OpenViking different from plain RAG?

OpenViking stores context as a directory tree rather than a flat vector pool. Searches can be scoped to a subtree, and each directory carries generated summaries so agents check relevance before reading full content.

Can I see what OpenViking stores about me?

Yes. OpenViking extracts memories as Markdown files under viking:// that you can open, inspect, edit, and merge.

Which agents does OpenViking work with?

OpenViking's integrations cover Claude Code, Codex, Cursor, TRAE, OpenClaw, Hermes, OpenCode, pi, DeerFlow, DSH, Doubao Work, LangChain, and general MCP clients.

What is the difference between find and search in OpenViking?

OpenViking's find runs a query directly, while search plans retrieval from session context.

Can I try OpenViking without installing it?

Yes. OpenViking Studio runs in the browser with no installation required.

Sources

  1. volcengine/OpenViking on GitHub
  2. OpenViking README
  3. openviking on PyPI
  4. OpenViking website
  5. Release v0.5.1.1 (Oct 20, 2026)
  6. Release [email protected] (Oct 19, 2026)
  7. Release [email protected] (Oct 19, 2026)
  8. Devtoolwatch Board, the week of Oct 25–31, 2026

OpenViking README badge

Maintainers can show the tool's weekly Board rank in their README:

Devtoolwatch badge for OpenViking

[![Devtoolwatch](https://index.devtoolwatch.com/badge/volcengine-openviking.svg)](https://index.devtoolwatch.com/tools/volcengine-openviking/)

How this page is made: Devtoolwatch writes each review from the project's README, releases, package metadata and website, with an AI model, then checks every claim against the source it cites and drops what the sources don't support. The numbers come from Devtoolwatch's daily GitHub snapshots, GitHub's stargazer history, and the npm and PyPI download APIs. Spotted an error? Email [email protected].