volcengine/OpenVikingPublic

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

AI summary: A self-evolving context database that unifies memory, RAG, and tool skills for advanced AI agents.

Stars
42.5K
+220 today
Forks
3.5K
Watchers
112
Open issues
482
Open PRs
941
Contributors
~340
Commits
3.3K
Branches
254

PythonAGPL-3.0Created Jan 5, 2026Last push todayLatest release [email protected]+381 stars this week+2K this month

Quick answers

What is OpenViking?
A self-evolving context database that unifies memory, RAG, and tool skills for advanced AI agents.
What does OpenViking do?
OpenViking provides a robust, centralized database architecture specifically designed to manage the context and state of autonomous AI agents. It solves the critical problem of agent amnesia and fragmented knowledge by providing a single interface for managing long-term memory, retrieving external knowledge (RAG), and registering callable skills. The technical foundation relies on a scalable storage backend that dynamically updates and refines the agent's context based on ongoing interactions, effectively allowing the agent to 'self-evolve' its understanding over time. What makes it distinctive is its holistic approach—rather than treating memory, search, and tools as separate modules, it unifies them into a cohesive context engine. It dramatically simplifies the development of complex, persistent agentic systems.
Who is OpenViking for?
This database is built for AI developers, backend engineers, and enterprise architects creating stateful, long-running agent applications. It requires a strong understanding of database systems, API design, and agentic workflows.
How do I get started with OpenViking?
git clone https://github.com/volcengine/OpenViking.git && docker-compose up -d
How popular is OpenViking on GitHub?
volcengine/OpenViking has 42,533 stars and 3,533 forks on GitHub, and gained 381 stars in the last 7 days.
What license does OpenViking use?
volcengine/OpenViking is released under the AGPL-3.0 license.

Star history

since Jul 29, 2026
020K40KJul 2026Sep 2026Oct 2026Nov 2026
42.5K stars as of Nov 20, 2026. Measured daily since Jul 29, 2026; GitHub no longer exposes earlier star timestamps.

Update history

5 recorded
  • Nov 13, 2026New release [email protected] (previously v0.5.3).
  • Nov 2, 2026New release v0.5.3 (previously v0.5.2).
  • Oct 24, 2026New release v0.5.2 (previously v0.5.0).
  • Oct 16, 2026New release v0.5.0 (previously v0.4.23).
  • Oct 6, 2026New release v0.4.23 (previously v0.4.21).

Contribution activity

commits per day, last 52 weeks
NovDecJanFebMarAprMayJunJulAugSepOctNovMonWedFri2025-11-16: 0 commits2025-11-17: 0 commits2025-11-18: 0 commits2025-11-19: 0 commits2025-11-20: 0 commits2025-11-21: 0 commits2025-11-22: 0 commits2025-11-23: 0 commits2025-11-24: 0 commits2025-11-25: 0 commits2025-11-26: 0 commits2025-11-27: 0 commits2025-11-28: 0 commits2025-11-29: 0 commits2025-11-30: 0 commits2025-12-01: 0 commits2025-12-02: 0 commits2025-12-03: 0 commits2025-12-04: 0 commits2025-12-05: 0 commits2025-12-06: 0 commits2025-12-07: 0 commits2025-12-08: 0 commits2025-12-09: 0 commits2025-12-10: 0 commits2025-12-11: 0 commits2025-12-12: 0 commits2025-12-13: 0 commits2025-12-14: 0 commits2025-12-15: 0 commits2025-12-16: 0 commits2025-12-17: 0 commits2025-12-18: 0 commits2025-12-19: 0 commits2025-12-20: 0 commits2025-12-21: 0 commits2025-12-22: 0 commits2025-12-23: 0 commits2025-12-24: 0 commits2025-12-25: 0 commits2025-12-26: 0 commits2025-12-27: 0 commits2025-12-28: 0 commits2025-12-29: 0 commits2025-12-30: 0 commits2025-12-31: 0 commits2026-01-01: 0 commits2026-01-02: 0 commits2026-01-03: 0 commits2026-01-04: 0 commits2026-01-05: 0 commits2026-01-06: 0 commits2026-01-07: 0 commits2026-01-08: 0 commits2026-01-09: 0 commits2026-01-10: 0 commits2026-01-11: 0 commits2026-01-12: 0 commits2026-01-13: 0 commits2026-01-14: 0 commits2026-01-15: 0 commits2026-01-16: 0 commits2026-01-17: 0 commits2026-01-18: 0 commits2026-01-19: 0 commits2026-01-20: 0 commits2026-01-21: 0 commits2026-01-22: 0 commits2026-01-23: 0 commits2026-01-24: 0 commits2026-01-25: 0 commits2026-01-26: 0 commits2026-01-27: 0 commits2026-01-28: 0 commits2026-01-29: 10 commits2026-01-30: 6 commits2026-01-31: 3 commits2026-02-01: 0 commits2026-02-02: 4 commits2026-02-03: 4 commits2026-02-04: 8 commits2026-02-05: 11 commits2026-02-06: 10 commits2026-02-07: 5 commits2026-02-08: 1 commit2026-02-09: 13 commits2026-02-10: 7 commits2026-02-11: 7 commits2026-02-12: 9 commits2026-02-13: 13 commits2026-02-14: 18 commits2026-02-15: 0 commits2026-02-16: 4 commits2026-02-17: 3 commits2026-02-18: 6 commits2026-02-19: 3 commits2026-02-20: 10 commits2026-02-21: 4 commits2026-02-22: 8 commits2026-02-23: 6 commits2026-02-24: 6 commits2026-02-25: 14 commits2026-02-26: 9 commits2026-02-27: 16 commits2026-02-28: 20 commits2026-03-01: 1 commit2026-03-02: 5 commits2026-03-03: 9 commits2026-03-04: 12 commits2026-03-05: 23 commits2026-03-06: 11 commits2026-03-07: 0 commits2026-03-08: 4 commits2026-03-09: 12 commits2026-03-10: 13 commits2026-03-11: 2 commits2026-03-12: 11 commits2026-03-13: 14 commits2026-03-14: 12 commits2026-03-15: 8 commits2026-03-16: 17 commits2026-03-17: 18 commits2026-03-18: 12 commits2026-03-19: 25 commits2026-03-20: 20 commits2026-03-21: 11 commits2026-03-22: 10 commits2026-03-23: 15 commits2026-03-24: 24 commits2026-03-25: 13 commits2026-03-26: 13 commits2026-03-27: 18 commits2026-03-28: 5 commits2026-03-29: 4 commits2026-03-30: 10 commits2026-03-31: 26 commits2026-04-01: 10 commits2026-04-02: 14 commits2026-04-03: 16 commits2026-04-04: 4 commits2026-04-05: 9 commits2026-04-06: 4 commits2026-04-07: 8 commits2026-04-08: 19 commits2026-04-09: 17 commits2026-04-10: 9 commits2026-04-11: 2 commits2026-04-12: 3 commits2026-04-13: 12 commits2026-04-14: 22 commits2026-04-15: 16 commits2026-04-16: 14 commits2026-04-17: 13 commits2026-04-18: 3 commits2026-04-19: 6 commits2026-04-20: 12 commits2026-04-21: 11 commits2026-04-22: 18 commits2026-04-23: 11 commits2026-04-24: 14 commits2026-04-25: 8 commits2026-04-26: 0 commits2026-04-27: 19 commits2026-04-28: 21 commits2026-04-29: 18 commits2026-04-30: 10 commits2026-05-01: 2 commits2026-05-02: 1 commit2026-05-03: 1 commit2026-05-04: 6 commits2026-05-05: 0 commits2026-05-06: 15 commits2026-05-07: 20 commits2026-05-08: 19 commits2026-05-09: 14 commits2026-05-10: 1 commit2026-05-11: 13 commits2026-05-12: 21 commits2026-05-13: 18 commits2026-05-14: 21 commits2026-05-15: 19 commits2026-05-16: 0 commits2026-05-17: 2 commits2026-05-18: 13 commits2026-05-19: 5 commits2026-05-20: 9 commits2026-05-21: 21 commits2026-05-22: 12 commits2026-05-23: 2 commits2026-05-24: 2 commits2026-05-25: 18 commits2026-05-26: 9 commits2026-05-27: 11 commits2026-05-28: 12 commits2026-05-29: 15 commits2026-05-30: 2 commits2026-05-31: 4 commits2026-06-01: 17 commits2026-06-02: 10 commits2026-06-03: 10 commits2026-06-04: 18 commits2026-06-05: 17 commits2026-06-06: 3 commits2026-06-07: 0 commits2026-06-08: 24 commits2026-06-09: 8 commits2026-06-10: 16 commits2026-06-11: 8 commits2026-06-12: 5 commits2026-06-13: 6 commits2026-06-14: 2 commits2026-06-15: 15 commits2026-06-16: 13 commits2026-06-17: 22 commits2026-06-18: 10 commits2026-06-19: 3 commits2026-06-20: 5 commits2026-06-21: 0 commits2026-06-22: 14 commits2026-06-23: 14 commits2026-06-24: 17 commits2026-06-25: 12 commits2026-06-26: 12 commits2026-06-27: 0 commits2026-06-28: 0 commits2026-06-29: 7 commits2026-06-30: 12 commits2026-07-01: 10 commits2026-07-02: 23 commits2026-07-03: 11 commits2026-07-04: 0 commits2026-07-05: 0 commits2026-07-06: 9 commits2026-07-07: 15 commits2026-07-08: 17 commits2026-07-09: 6 commits2026-07-10: 16 commits2026-07-11: 7 commits2026-07-12: 0 commits2026-07-13: 25 commits2026-07-14: 11 commits2026-07-15: 11 commits2026-07-16: 13 commits2026-07-17: 4 commits2026-07-18: 1 commit2026-07-19: 3 commits2026-07-20: 9 commits2026-07-21: 5 commits2026-07-22: 15 commits2026-07-23: 9 commits2026-07-24: 16 commits2026-07-25: 3 commits2026-07-26: 3 commits2026-07-27: 9 commits2026-07-28: 12 commits2026-07-29: 12 commits2026-07-30: 14 commits2026-07-31: 15 commits2026-08-01: 1 commit2026-08-02: 0 commits2026-08-03: 11 commits2026-08-04: 7 commits2026-08-05: 9 commits2026-08-06: 12 commits2026-08-07: 16 commits2026-08-08: 3 commits2026-08-09: 0 commits2026-08-10: 20 commits2026-08-11: 8 commits2026-08-12: 12 commits2026-08-13: 10 commits2026-08-14: 10 commits2026-08-15: 1 commit2026-08-16: 0 commits2026-08-17: 6 commits2026-08-18: 24 commits2026-08-19: 13 commits2026-08-20: 13 commits2026-08-21: 12 commits2026-08-22: 2 commits2026-08-23: 0 commits2026-08-24: 26 commits2026-08-25: 15 commits2026-08-26: 24 commits2026-08-27: 11 commits2026-08-28: 15 commits2026-08-29: 2 commits2026-08-30: 2 commits2026-08-31: 24 commits2026-09-01: 14 commits2026-09-02: 13 commits2026-09-03: 9 commits2026-09-04: 25 commits2026-09-05: 2 commits2026-09-06: 1 commit2026-09-07: 16 commits2026-09-08: 24 commits2026-09-09: 16 commits2026-09-10: 14 commits2026-09-11: 17 commits2026-09-12: 5 commits2026-09-13: 2 commits2026-09-14: 20 commits2026-09-15: 23 commits2026-09-16: 17 commits2026-09-17: 23 commits2026-09-18: 17 commits2026-09-19: 9 commits2026-09-20: 16 commits2026-09-21: 26 commits2026-09-22: 24 commits2026-09-23: 19 commits2026-09-24: 28 commits2026-09-25: 11 commits2026-09-26: 1 commit2026-09-27: 0 commits2026-09-28: 15 commits2026-09-29: 25 commits2026-09-30: 16 commits2026-10-01: 16 commits2026-10-02: 4 commits2026-10-03: 10 commits2026-10-04: 0 commits2026-10-05: 13 commits2026-10-06: 15 commits2026-10-07: 13 commits2026-10-08: 15 commits2026-10-09: 8 commits2026-10-10: 7 commits2026-10-11: 6 commits2026-10-12: 21 commits2026-10-13: 27 commits2026-10-14: 16 commits2026-10-15: 26 commits2026-10-16: 20 commits2026-10-17: 1 commit2026-10-18: 1 commit2026-10-19: 26 commits2026-10-20: 32 commits2026-10-21: 16 commits2026-10-22: 11 commits2026-10-23: 7 commits2026-10-24: 1 commit2026-10-25: 0 commits2026-10-26: 18 commits2026-10-27: 20 commits2026-10-28: 24 commits2026-10-29: 25 commits2026-10-30: 21 commits2026-10-31: 3 commits2026-11-01: 4 commits2026-11-02: 27 commits2026-11-03: 8 commits2026-11-04: 27 commits2026-11-05: 12 commits2026-11-06: 12 commits2026-11-07: 3 commits2026-11-08: 3 commits2026-11-09: 17 commits2026-11-10: 34 commits2026-11-11: 22 commits2026-11-12: 29 commits2026-11-13: 23 commits2026-11-14: 0 commits
3,272 commits in the last yearLessMore

Signals and awards

derived from tracked data
  • Widely adopted

    42,533 stars

  • Very active

    3,272 commits in 52 weeks

  • Community-driven

    ~340 contributors

  • Repeat trending

    40 trending appearances

What OpenViking does

OpenViking provides a robust, centralized database architecture specifically designed to manage the context and state of autonomous AI agents. It solves the critical problem of agent amnesia and fragmented knowledge by providing a single interface for managing long-term memory, retrieving external knowledge (RAG), and registering callable skills. The technical foundation relies on a scalable storage backend that dynamically updates and refines the agent's context based on ongoing interactions, effectively allowing the agent to 'self-evolve' its understanding over time. What makes it distinctive is its holistic approach—rather than treating memory, search, and tools as separate modules, it unifies them into a cohesive context engine. It dramatically simplifies the development of complex, persistent agentic systems.

This database is built for AI developers, backend engineers, and enterprise architects creating stateful, long-running agent applications. It requires a strong understanding of database systems, API design, and agentic workflows.

  • Unified Context Management: Consolidates agent memory, document retrieval, and tool execution into a single, queryable database layer.
  • Self-Evolving Memory: Automatically refines and updates stored context based on user interactions, preventing the buildup of irrelevant data.
  • Integrated Skill Registry: Allows developers to register and manage executable plugins directly within the database architecture.
  • Scalable Storage Backend: Engineered to handle high-throughput read/write operations required by multi-agent enterprise deployments.
  • Multi-Tenant Architecture: Securely isolates context and memory state across different users and agent instances.

Where teams use it

Persistent AI Companions

Developers build highly personalized consumer chatbots that remember detailed user preferences and past conversations across months of usage.

Enterprise Knowledge Agents

Organizations deploy internal support bots that combine company documentation (RAG) with the ability to execute API calls (Skills) seamlessly.

Multi-Agent Orchestration

System architects use the unified database to share context safely and efficiently between different specialized agents in a complex workflow.

Dynamic Data Integration

Data scientists connect real-time data streams to the context database, ensuring agents always have access to the latest market or operational metrics.

Getting started: git clone https://github.com/volcengine/OpenViking.git && docker-compose up -d

README

main branch
OpenViking

The Context Database for AI Agents

English / 中文 / 日本語

Website · Live Demo · GitHub · Issues · Docs · Blog

release stars issues contributors license last commit

Deploy on Railway

Lark Lark · WeChat WeChat · Discord Discord · X X

volcengine%2FOpenViking | Trendshift


What is OpenViking

OpenViking is an open-source context database for AI agents — one filesystem for everything an agent knows: knowledge, memory, and skills.

Most agent memory is a black box: text goes in, embeddings come out, and nobody can see what was actually stored. OpenViking organizes context as a virtual filesystem under viking:// instead. Agents navigate it like files — ls, tree, read, write, grep — and you can open any directory to inspect and edit what your agent knows. Every directory carries a generated summary, so agents can scan summaries first and decide what to read.

OpenViking Studio Filesystem: browse personal memories and resources

Try OpenViking Studio in your browser, no installation required. Self-host Web Studio.

Why OpenViking

  • One filesystem for knowledge, memory, and skills. Resources hold documents and code; memories retain user preferences and experience; skills define how to perform tasks — not just extracted facts, but the full context, each with a viking:// URI for browsing and retrieval. → Viking URI · Context types
  • Search a directory, not the whole index. Scope semantic search to a project or memory subtree instead of scanning a flat vector pool. find runs a query directly; search plans retrieval from session context. → Retrieval
  • Read the summary before the source. Generated directory abstracts (L0) and overviews (L1) let agents judge relevance before opening full content (L2). → Context layers
  • Sessions become files you can read. Committing a session archives the conversation and extracts memories as Markdown you can inspect, edit, and merge. With VikingBot enabled, ov compile organizes source material into a wiki, knowledge graph, or report. → Sessions · Context compilation

Architecture · Design rationale

viking://
├── resources/              # Resources: project docs, repos, web pages, etc.
│   └── my_project/
│       ├── docs/
│       │   ├── api/
│       │   └── tutorials/
│       └── src/
└── user/
    └── {user_id}/
        ├── memories/
        │   └── preferences/
        │       ├── writing_style
        │       └── coding_habits
        ├── resources/
        │   └── private_project/
        ├── skills/
        │   ├── search_code
        │   └── analyze_data
        └── peers/
            └── web-visitor-alice/

The three loading tiers:

  • L0 (Abstract): a one-sentence summary for quick relevance checks.
  • L1 (Overview): core information and usage scenarios for planning.
  • L2 (Details): the full original data, read only when needed.

Semantically processed directories carry L0/L1 summaries, so agents can judge relevance before reading full files:

viking://resources/my_project/
├── .abstract.md           # L0: quick relevance check
├── .overview.md           # L1: structure and key points
└── docs/
    ├── .abstract.md
    ├── .overview.md
    └── api/
        ├── auth.md         # L2: full content, loaded on demand
        └── endpoints.md

Proof it works

OpenViking 0.3.22 has been evaluated on long-conversation user memory (LoCoMo) and multi-turn agent tasks (tau2-bench). Full results and setup details, including knowledge-base QA, are in the benchmark report; reproduction scripts live in ./benchmark.

To run your own experiments, follow the RAG and Memory evaluation guide for routine VikingBot memory evaluation, published LoCoMo reproduction, separate mem0 prompt comparisons, and RAG evaluation.

The memory evaluation used Doubao 2.0 Pro as the VLM and Doubao-embedding-vision-251215 as the embedding model.

Benchmark results. LoCoMo accuracy: OpenClaw 24.20% native vs 82.08% with OpenViking; Hermes 33.38% vs 82.86%; Claude Code 57.21% vs 80.32%. tau2-bench task success: Retail 70.94% vs 77.81%; Airline 54.38% vs 66.25%.
  • User memory (LoCoMo): with OpenViking, all three agent integrations land at 80–83% accuracy — up from 24–57% on their native memory — while input tokens drop by 34.3–91.0% and query latency by 58.45–66.10%.
  • Agent experience (tau2-bench): experience memory lifts task success by +6.87pp (retail) and +11.87pp (airline) over the same LLM without memory.

Quick start

Set up an OpenViking server first; if you already have one, skip to Use it with your agent. Deploying your own needs uv, Python 3.10+ and a model provider with an embedding model and a VLM.

Let your agent deploy it
Follow this guide to install and start an OpenViking Server for me:

https://docs.openviking.ai/en/getting-started/04-setup-for-agent

Ask me for the model provider, the models, the workspace directory and whether
other machines need to reach the server; don't guess. When you ask for the model
API key, tell me how to hand it over if I'd rather not paste it into this chat,
and never repeat it back.

When it's running, tell me the server's address and whether auth is turned on.

Your agent will first ask which model provider to use and for its API key.

Deploy it yourself

Install OpenViking and run the setup wizard, which configures the models:

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

init writes ~/.openviking/ov.conf and supports Volcengine, OpenAI, Codex OAuth, Kimi, GLM and local Ollama; see the configuration guide. The server runs in the foreground, so keep this terminal open.

Use OpenViking Service (hosted by Volcengine)

The same OpenViking service, run for you by Volcengine. The first 50 files are free. Activate it on the Volcengine product page, then create an API key in the console under User Management → API Key. The server address is https://api.vikingdb.cn-beijing.volces.com/openviking; you'll need it and the API key when you connect your agent.

Explore with the CLI

The openviking package includes the ov CLI. With the server running, import a repository and search it:

ov status
ov add-resource https://github.com/volcengine/OpenViking
# Replace TASK_ID with the returned task_id; repeat until status is completed
ov task status TASK_ID
ov ls viking://resources/
ov tree viking://resources/volcengine -L 2
ov find "what is openviking"
ov grep "openviking" --uri viking://resources/volcengine/OpenViking/docs/en

ov find returns matching context with URIs you can inspect. For client configuration (ov config), standalone CLI installs, and index maintenance, see CLI setup.

Build your own integration with the Python, Go, or TypeScript SDK, or the HTTP API.

Use it with your agent

Connect your coding agent to OpenViking for cross-session memory. The memory plugin installer covers Claude Code, Codex, Cursor, TRAE, OpenCode and more, and detects which ones you have.

Let your agent install it
Install the OpenViking memory plugin and help me start using it. OpenViking is a context database for AI agents (https://github.com/volcengine/OpenViking); the command and link below are from its official docs.

1. Run the official installer. It needs network access and writes to my home directory; if your sandbox blocks that, ask me for approval:
   curl -fsSL https://openviking.ai/install | bash -s -- --yes --lang en --url <SERVER_URL> --api-key <API_KEY>
2. Once it's installed, read the OpenViking onboarding guide and use it to walk me through setup. The guide has you ask me before choices like importing history or adding a working rule. docs.openviking.ai and docs.openviking.net are the same docs site; use whichever loads:
   https://docs.openviking.ai/agents/en/onboarding.md

The API key is a user key I created for myself on that server; the installer stores it in ~/.openviking/ovcli.conf on this machine. Use it in the command as it is; there's no need to show it in your replies, and don't write it into any other file or rule. Talk to me in English.

Before you send it, replace <SERVER_URL> with your server's address and <API_KEY> with a user key; a root key can't read or write memories. If the server has auth turned off, use --api-key ''.

Install it yourself

Run the installer, then pick the tools and a server when asked: Self-hosted / local for a server on this machine, Volcengine OpenViking Cloud for OpenViking Service, or Custom URL for any other address. If auth is on, enter a user key, not the root key.

curl -fsSL https://openviking.ai/install | bash
# AI agents: not sure about running this? The script's header says what it does and how to verify it.

Then restart your agent. The installer ends with the next steps for each tool, for example:

  Claude Code
    Next: restart Claude Code
    ⋮
    Verify: run /openviking-memory:ov in a session

On its first start Codex stops at Hooks need review; choose Trust all and continue.

Try it: ask it to remember one of your preferences, then ask about it in a new session a little later. Memories are processed in the background, so it's normal not to find one right after you say it.

The installer needs macOS or Linux, Node.js 18+ and curl, no sudo. On Windows, use the desktop app.

Setup guides for each integration:


Claude

Hooks + MCP

Codex

Hooks + MCP

Cursor

Hooks + MCP

TRAE

Hooks + MCP

OpenClaw

Context engine

Hermes

Built-in

OpenCode

Plugin + MCP

pi

Native extension

DeerFlow

Plugin + MCP

DSH

Plugin + MCP

Doubao Work

Connector

LangChain

Tools + store

General integrations


Agent Plugins 1.0

MCP clients

For setup instructions and integration details, see Integrations.

For framework applications, see the LangChain/LangGraph and Microsoft Agent Framework packages, with a guide to adding other adapters.

Desktop App (Beta)

The desktop app is a console for macOS and Windows x64 (beta). It configures supported local agent integrations, inspects recall and capture events in sessions, and syncs local memories and skills to OpenViking.

Download:

VikingBot

VikingBot is an AI agent framework built on top of OpenViking:

pip install "openviking[bot]"
openviking-server --with-bot
ov chat   # in another terminal

The official Docker image bundles VikingBot and starts it by default alongside the server and console UI. Details: VikingBot guide.

Deploy in production

Run the open-source server in your own environment under AGPLv3. It requires no activation key. Start with server setup or the Docker and deployment guide.

The server supports accounts and user isolation and opt-in resource ACLs. Configure authentication before exposing it beyond localhost.

Commercial editions

Managed SaaS

☁️ Managed SaaS

Volcano Engine hosts and operates OpenViking. Personal and Enterprise plans cover individual and team use, with migration tooling for open-source deployments. See the service documentation for plans and limits. Hosting outside China is planned on BytePlus.

Self-Managed

🏢 Self-Managed

Deploy in your own cloud account / VPC (BYOC) or an offline environment. This edition adds distributed deployment and official support, activated by a license key. Contact the team.

Research

Memory that evolves with your agent. VikingMem develops an event-driven approach to extracting, updating, and consolidating long-term memory, giving stateful agents a way to retain useful experience as interactions accumulate. OpenViking open-sources a subset of these core capabilities.

VikingMem: A Memory Base Management System for Stateful LLM-based Applications
Jiajie Fu, Junwen Chen, Mengzhao Wang, Aoxiang He, Maojia Sheng, Xiangyu Ke, Yifan Zhu, and Yunjun Gao.
arXiv:2605.29640, 2026. Presented at VLDB 2026 in September.
📄 Read the paper on arXiv · Read PDF

Directory structure as retrieval context. This paper provides the formal foundations, index design, and experimental evidence behind OpenViking’s directory-aware retrieval. It defines directory-scoped query and maintenance operations and introduces TrieHI, which OpenViking integrates to resolve directory scopes before vector ranking. This connects the filesystem paradigm to retrieval: agents can search a project or memory subtree, retain its surrounding context, and reorganize it as knowledge evolves.

Directory-Aware Query and Maintenance in Vector Databases
Mengzhao Wang, Zheng Gong, Jingpei Hu, Jiajie Fu, Maojia Sheng, Junwen Chen, and Yifan Zhu.
arXiv:2606.16903, 2026. Accepted by ICDE.
📄 Read the paper on arXiv · Read PDF

Retrieve the evidence you need with fewer tokens. VikingRAG combines semantic search with document structure, exposing relevant directory segments as evidence gaps arise. Its core mechanisms are integrated into OpenViking. The paper further explores reusing retrieval traces and escalating to multi-round retrieval only when needed, reducing repeated exploration while preserving answer quality.

VikingRAG: Accurate and Token-efficient Retrieval-augmented Generation over Structured Documents
Peiyuan Gao, Gaoyuan Zhang, Haojie Qin, Yahui Sun, Qianyi Zhang, Yunhao Zhang, Zeyu Wang, and Wei Lu.
arXiv:2609.11390, 2026. Submitted.
📄 Read the paper on arXiv · Read PDF

Partner Projects

  • deer-flow - Open-source long-horizon SuperAgent harness
  • NoKV - AI native distributed file system
  • loopx - Lightweight loop engineering state kernel
  • Hermes Agent - The agent that grows with you
  • NVIDIA cuVS - GPU-accelerated vector search and clustering library

To propose a partnership, open an issue.

Community & Contributing

OpenViking contributors

Security and privacy

For vulnerability reporting and supported versions, see SECURITY.md

License

Copyright © 2026 Volcano Engine (火山引擎) and the OpenViking contributors.

The OpenViking project uses different licenses for different components:

  • Main Project: AGPLv3 - see the LICENSE file for details
  • crates/ov_cli: Apache 2.0 - see the LICENSE for details
  • examples: Apache 2.0 - see the LICENSE for details. The Hermes plugin in examples/hermes-plugin retains its MIT license.
  • integrations/frameworks: the LangChain/LangGraph adapter uses AGPLv3 and the Microsoft Agent Framework adapter uses MIT; see each package's LICENSE. Their runnable examples retain Apache 2.0; see each package's examples/LICENSE.
  • third_party: Respective original licenses of third-party projects

Please also note: any wholesale rewriting and subsequent re-open-sourcing of this project is prohibited; any request to translate all or a substantial portion of the project code into another coding language requires prior written authorization from Volcano Engine. For permission requests, please contact the Volcano Engine team — see Community & Contributing for contact channels.

View on GitHub

Recent activity

commits and pull requests

Releases and announcements

113 total
  1. ## 更新 - 支持上下文使用次数与最近使用时间标签更新 - 支持批量设置上下文标签 - 支持 Python 同步与异步客户端调用 `batch_set_tags`

  2. v0.5.4.1v0.5.4.1Nov 11, 2026

    ## 中文 ### 修复与改进 - **资源 write**:修正文件处理与父目录刷新的衔接,保留每个文件自身的摘要和索引工作;父目录独立刷新,不替代尚未完成的文件索引。`wait=true` 等待文件自身的处理完成,不等待父目录 L0/L1 和目录索引刷新,请求参数不变。[#7383](https://github.com/volcengine/OpenViking/pull/7383) - **目录摘要**:待刷新期间复用已有文件摘要;恢复 `semantic.overview_sample_limit` 默认 32 项的稳定采样,文件自身仍完整建立索引。[#7380](https://github.com/volcengine/OpenViking/pull/7380)、[#7387](https://github.com/volcengine/OpenViking/pull/7387) - **存储稳定性**:修复 `stat` 按记录 ID 查询的兼容性、向量 Collection 生命周期,以及多写元数据不完整时的启动问题。[#7377](https://github.com/volcengine/OpenViking/pull/7377)、[#7355](https://github.com/volcengine/OpenViking/pull/7355)、[#7381](https://github.com/volcengine/OpenViking/pull/7381) - **Connector 与流式响应**:支持带作用域的 OAuth 引用,原生导入保留 Connector Watch;修复 Gateway 流式错误的结束标记。[#7259](https://github.com/volcengine/OpenViking/pull/7259)、[#7364](https://github.com/volcengine/OpenViking/pull/7364)、[#7010](https://github.com/volcengine/OpenViking/pull/7010) ### 其他更新 新增多写 backup 动态切换、账户管理的 global 系统 Skill 目录、火山引擎浏览器授权配置,并更新 Console、FileBrowser、Wiki 和 Agent 体验界面。详见[完整变更](https://github.com/volcengine/OpenViking/compare/v0.5.4...v0.5.4.1)。 ### 升级说明 ```bash pip install --upgrade 'openviking==0.5.4.1' ``` - 目录 `stat` / `ls` 的 `count

  3. ## Install macOS / Linux / WSL2 (this version): ```sh curl -fsSL https://ovrelease.tos-cn-beijing.volces.com/superagent/releases/0.5.2-preview/install.sh | sh -s -- --version 0.5.2-preview ``` Windows users: use the PowerShell one-line installer in the shared guide below. Windows previews are published separately; this command and these release assets are for macOS / Linux. [Installation and setup guide](https://github.com/volcengine/OpenViking/blob/feat/vika/superagent/install.md)

  4. ## Install macOS / Linux / WSL2 (this version): ```sh curl -fsSL https://ovrelease.tos-cn-beijing.volces.com/superagent/releases/0.5.1-preview/install.sh | sh -s -- --version 0.5.1-preview ``` Windows users: use the PowerShell one-line installer in the shared guide below. Windows previews are published separately; this command and these release assets are for macOS / Linux. [Installation and setup guide](https://github.com/volcengine/OpenViking/blob/feat/vika/superagent/install.md)

  5. v0.5.4v0.5.4Nov 9, 2026

    ## 中文 ### 主要更新 - **MCP 与实时上下文**:注册 Streamable HTTP MCP Server,检索、调用工具,并把工具挂载为可读取的 Markdown 节点。提供 REST、CLI 和 SDK 入口;同步上游工具后可查看受影响的挂载。[#6263](https://github.com/volcengine/OpenViking/pull/6263) - **Agent 规则查询**:多种 Agent 集成增加 Rule Search 入口,在执行操作前按需查找适用规则。升级插件或重连 MCP 客户端后生效;这不会自动增加每轮拦截。[#7005](https://github.com/volcengine/OpenViking/pull/7005) - **Watch 多节点运行**:增加 MySQL Watch 存储和列表分页;Connector 授权撤销后自动暂停,重新授权时只恢复此前自动暂停的任务。[#6756](https://github.com/volcengine/OpenViking/pull/6756)、[#6944](https://github.com/volcengine/OpenViking/pull/6944) - **会话恢复与记忆整理**:失败或失去执行任务的归档支持带内容哈希的显式重试;可用 `ov compile memory` 整理当前用户的已有记忆。[#5802](https://github.com/volcengine/OpenViking/pull/5802)、[#6829](https://github.com/volcengine/OpenViking/pull/6829) - **配置与访问控制**:Account 可独立配置抽取开关、生成语言及解析凭证;普通用户可管理自己的添加位置和记忆策略。新增由 ROOT 管理的 GLOBAL 身份及按数据所有者配置的读取策略。[#6986](https://github.com/volcengine/OpenViking/pull/6986)、[#7050](https://github.com/volcengine/OpenViking/pull/7050)、[#6900](https://github.com/volcengine/OpenViking/pull/6900)、[#6932](https://github.com/volcengine/OpenViking/pull/6932) ### 可选可观测性上报 新增可选 Kafka 事件与存量盘点上报,需要安装 `openviking[reporting]` 并显式开启,默认关闭。上报保留部署、数据空间、用户和模型费用归属,不包含原文、提示词或 API Key;采用内

Code frequency

additions and deletions
+242.8K-242.8KWeek of 2026-01-25: +242,772 linesWeek of 2026-01-25: -722 linesWeek of 2026-02-01: +25,948 linesWeek of 2026-02-01: -7,733 linesWeek of 2026-02-08: +47,852 linesWeek of 2026-02-08: -23,704 linesWeek of 2026-02-15: +10,858 linesWeek of 2026-02-15: -1,441 linesWeek of 2026-02-22: +79,035 linesWeek of 2026-02-22: -23,778 linesWeek of 2026-03-01: +17,818 linesWeek of 2026-03-01: -10,314 linesWeek of 2026-03-08: +33,436 linesWeek of 2026-03-08: -13,152 linesWeek of 2026-03-15: +47,138 linesWeek of 2026-03-15: -16,762 linesWeek of 2026-03-22: +63,712 linesWeek of 2026-03-22: -11,733 linesWeek of 2026-03-29: +53,379 linesWeek of 2026-03-29: -25,255 linesWeek of 2026-04-05: +32,078 linesWeek of 2026-04-05: -76,007 linesWeek of 2026-04-12: +58,701 linesWeek of 2026-04-12: -8,806 linesWeek of 2026-04-19: +31,337 linesWeek of 2026-04-19: -5,949 linesWeek of 2026-04-26: +28,115 linesWeek of 2026-04-26: -11,006 linesWeek of 2026-05-03: +45,353 linesWeek of 2026-05-03: -10,686 linesWeek of 2026-05-10: +54,940 linesWeek of 2026-05-10: -15,605 linesWeek of 2026-05-17: +98,495 linesWeek of 2026-05-17: -15,120 linesWeek of 2026-05-24: +32,562 linesWeek of 2026-05-24: -24,618 linesWeek of 2026-05-31: +46,158 linesWeek of 2026-05-31: -27,715 linesWeek of 2026-06-07: +47,352 linesWeek of 2026-06-07: -11,664 linesWeek of 2026-06-14: +70,636 linesWeek of 2026-06-14: -13,684 linesWeek of 2026-06-21: +46,037 linesWeek of 2026-06-21: -10,883 linesWeek of 2026-06-28: +56,343 linesWeek of 2026-06-28: -4,207 linesWeek of 2026-07-05: +58,663 linesWeek of 2026-07-05: -11,751 linesWeek of 2026-07-12: +27,733 linesWeek of 2026-07-12: -17,510 linesWeek of 2026-07-19: +37,246 linesWeek of 2026-07-19: -12,561 linesWeek of 2026-07-26: +73,344 linesWeek of 2026-07-26: -35,104 linesWeek of 2026-08-02: +44,688 linesWeek of 2026-08-02: -15,767 linesWeek of 2026-08-09: +45,074 linesWeek of 2026-08-09: -29,720 linesWeek of 2026-08-16: +40,067 linesWeek of 2026-08-16: -22,500 linesWeek of 2026-08-23: +29,780 linesWeek of 2026-08-23: -9,968 linesWeek of 2026-08-30: +51,459 linesWeek of 2026-08-30: -15,917 linesWeek of 2026-09-06: +54,537 linesWeek of 2026-09-06: -9,564 linesWeek of 2026-09-13: +72,327 linesWeek of 2026-09-13: -52,119 linesWeek of 2026-09-20: +101,781 linesWeek of 2026-09-20: -35,752 linesWeek of 2026-09-27: +39,274 linesWeek of 2026-09-27: -31,762 linesWeek of 2026-10-04: +121,751 linesWeek of 2026-10-04: -66,103 linesWeek of 2026-10-11: +96,694 linesWeek of 2026-10-11: -48,455 linesWeek of 2026-10-18: +56,052 linesWeek of 2026-10-18: -20,150 linesWeek of 2026-10-25: +188,523 linesWeek of 2026-10-25: -125,089 linesWeek of 2026-11-01: +112,120 linesWeek of 2026-11-01: -40,256 linesWeek of 2026-11-08: +172,490 linesWeek of 2026-11-08: -61,196 linesJan 25, 2026Nov 8, 2026
+2.7M lines added, -1M removed over the last year.

Commits per week

last 52 weeks
1280Week of 2025-11-16: 0 commitsWeek of 2025-11-23: 0 commitsWeek of 2025-11-30: 0 commitsWeek of 2025-12-07: 0 commitsWeek of 2025-12-14: 0 commitsWeek of 2025-12-21: 0 commitsWeek of 2025-12-28: 0 commitsWeek of 2026-01-04: 0 commitsWeek of 2026-01-11: 0 commitsWeek of 2026-01-18: 0 commitsWeek of 2026-01-25: 19 commitsWeek of 2026-02-01: 42 commitsWeek of 2026-02-08: 68 commitsWeek of 2026-02-15: 30 commitsWeek of 2026-02-22: 79 commitsWeek of 2026-03-01: 61 commitsWeek of 2026-03-08: 68 commitsWeek of 2026-03-15: 111 commitsWeek of 2026-03-22: 98 commitsWeek of 2026-03-29: 84 commitsWeek of 2026-04-05: 68 commitsWeek of 2026-04-12: 83 commitsWeek of 2026-04-19: 80 commitsWeek of 2026-04-26: 71 commitsWeek of 2026-05-03: 75 commitsWeek of 2026-05-10: 93 commitsWeek of 2026-05-17: 64 commitsWeek of 2026-05-24: 69 commitsWeek of 2026-05-31: 79 commitsWeek of 2026-06-07: 67 commitsWeek of 2026-06-14: 70 commitsWeek of 2026-06-21: 69 commitsWeek of 2026-06-28: 63 commitsWeek of 2026-07-05: 70 commitsWeek of 2026-07-12: 65 commitsWeek of 2026-07-19: 60 commitsWeek of 2026-07-26: 66 commitsWeek of 2026-08-02: 58 commitsWeek of 2026-08-09: 61 commitsWeek of 2026-08-16: 70 commitsWeek of 2026-08-23: 93 commitsWeek of 2026-08-30: 89 commitsWeek of 2026-09-06: 93 commitsWeek of 2026-09-13: 111 commitsWeek of 2026-09-20: 125 commitsWeek of 2026-09-27: 86 commitsWeek of 2026-10-04: 71 commitsWeek of 2026-10-11: 117 commitsWeek of 2026-10-18: 94 commitsWeek of 2026-10-25: 111 commitsWeek of 2026-11-01: 93 commitsWeek of 2026-11-08: 128 commitsNov 16, 2025Nov 8, 2026
3.3K commits in the last 52 weeks.

When work happens

weekday and hour
SunMonTueWedThuFriSat036912151821Sun 0:00 — 1 commitsSun 1:00 — 5 commitsSun 2:00 — 1 commitsSun 3:00 — 0 commitsSun 4:00 — 0 commitsSun 5:00 — 0 commitsSun 6:00 — 2 commitsSun 7:00 — 1 commitsSun 8:00 — 0 commitsSun 9:00 — 7 commitsSun 10:00 — 5 commitsSun 11:00 — 5 commitsSun 12:00 — 1 commitsSun 13:00 — 6 commitsSun 14:00 — 8 commitsSun 15:00 — 3 commitsSun 16:00 — 3 commitsSun 17:00 — 4 commitsSun 18:00 — 14 commitsSun 19:00 — 6 commitsSun 20:00 — 8 commitsSun 21:00 — 8 commitsSun 22:00 — 8 commitsSun 23:00 — 11 commitsMon 0:00 — 4 commitsMon 1:00 — 4 commitsMon 2:00 — 6 commitsMon 3:00 — 3 commitsMon 4:00 — 1 commitsMon 5:00 — 1 commitsMon 6:00 — 1 commitsMon 7:00 — 5 commitsMon 8:00 — 6 commitsMon 9:00 — 5 commitsMon 10:00 — 32 commitsMon 11:00 — 70 commitsMon 12:00 — 26 commitsMon 13:00 — 21 commitsMon 14:00 — 53 commitsMon 15:00 — 53 commitsMon 16:00 — 52 commitsMon 17:00 — 62 commitsMon 18:00 — 32 commitsMon 19:00 — 48 commitsMon 20:00 — 58 commitsMon 21:00 — 41 commitsMon 22:00 — 7 commitsMon 23:00 — 8 commitsTue 0:00 — 1 commitsTue 1:00 — 3 commitsTue 2:00 — 2 commitsTue 3:00 — 2 commitsTue 4:00 — 1 commitsTue 5:00 — 1 commitsTue 6:00 — 4 commitsTue 7:00 — 2 commitsTue 8:00 — 2 commitsTue 9:00 — 7 commitsTue 10:00 — 22 commitsTue 11:00 — 80 commitsTue 12:00 — 37 commitsTue 13:00 — 17 commitsTue 14:00 — 48 commitsTue 15:00 — 64 commitsTue 16:00 — 50 commitsTue 17:00 — 55 commitsTue 18:00 — 34 commitsTue 19:00 — 44 commitsTue 20:00 — 54 commitsTue 21:00 — 37 commitsTue 22:00 — 24 commitsTue 23:00 — 16 commitsWed 0:00 — 10 commitsWed 1:00 — 3 commitsWed 2:00 — 0 commitsWed 3:00 — 5 commitsWed 4:00 — 0 commitsWed 5:00 — 1 commitsWed 6:00 — 0 commitsWed 7:00 — 0 commitsWed 8:00 — 1 commitsWed 9:00 — 4 commitsWed 10:00 — 18 commitsWed 11:00 — 75 commitsWed 12:00 — 34 commitsWed 13:00 — 18 commitsWed 14:00 — 66 commitsWed 15:00 — 55 commitsWed 16:00 — 57 commitsWed 17:00 — 74 commitsWed 18:00 — 29 commitsWed 19:00 — 47 commitsWed 20:00 — 39 commitsWed 21:00 — 35 commitsWed 22:00 — 19 commitsWed 23:00 — 5 commitsThu 0:00 — 7 commitsThu 1:00 — 0 commitsThu 2:00 — 2 commitsThu 3:00 — 2 commitsThu 4:00 — 2 commitsThu 5:00 — 4 commitsThu 6:00 — 2 commitsThu 7:00 — 0 commitsThu 8:00 — 1 commitsThu 9:00 — 1 commitsThu 10:00 — 16 commitsThu 11:00 — 52 commitsThu 12:00 — 34 commitsThu 13:00 — 23 commitsThu 14:00 — 53 commitsThu 15:00 — 53 commitsThu 16:00 — 68 commitsThu 17:00 — 78 commitsThu 18:00 — 34 commitsThu 19:00 — 55 commitsThu 20:00 — 62 commitsThu 21:00 — 39 commitsThu 22:00 — 20 commitsThu 23:00 — 13 commitsFri 0:00 — 4 commitsFri 1:00 — 9 commitsFri 2:00 — 4 commitsFri 3:00 — 2 commitsFri 4:00 — 4 commitsFri 5:00 — 1 commitsFri 6:00 — 0 commitsFri 7:00 — 1 commitsFri 8:00 — 2 commitsFri 9:00 — 2 commitsFri 10:00 — 11 commitsFri 11:00 — 67 commitsFri 12:00 — 33 commitsFri 13:00 — 25 commitsFri 14:00 — 64 commitsFri 15:00 — 63 commitsFri 16:00 — 59 commitsFri 17:00 — 56 commitsFri 18:00 — 35 commitsFri 19:00 — 37 commitsFri 20:00 — 37 commitsFri 21:00 — 20 commitsFri 22:00 — 5 commitsFri 23:00 — 13 commitsSat 0:00 — 10 commitsSat 1:00 — 6 commitsSat 2:00 — 4 commitsSat 3:00 — 1 commitsSat 4:00 — 0 commitsSat 5:00 — 0 commitsSat 6:00 — 0 commitsSat 7:00 — 1 commitsSat 8:00 — 1 commitsSat 9:00 — 6 commitsSat 10:00 — 12 commitsSat 11:00 — 10 commitsSat 12:00 — 8 commitsSat 13:00 — 7 commitsSat 14:00 — 14 commitsSat 15:00 — 18 commitsSat 16:00 — 17 commitsSat 17:00 — 12 commitsSat 18:00 — 11 commitsSat 19:00 — 13 commitsSat 20:00 — 15 commitsSat 21:00 — 5 commitsSat 22:00 — 9 commitsSat 23:00 — 10 commits
Commit volume by weekday and hour (UTC). Larger dots mean more commits.
DateListRankStars gained
Sep 17, 2026monthly#13+9,114
Sep 16, 2026monthly#13+9,114
Sep 15, 2026monthly#10+8,921
Sep 14, 2026monthly#5+8,745
Sep 13, 2026monthly#5+8,682
Sep 12, 2026monthly#6+8,596
Sep 11, 2026monthly#7+8,469
Sep 10, 2026monthly#6+8,284
Sep 9, 2026monthly#6+8,115
Sep 8, 2026monthly#6+7,993
Sep 7, 2026monthly#5+7,895
Sep 6, 2026monthly#7+7,811
Sep 5, 2026monthly#6+7,710
Sep 4, 2026monthly#6+7,659
Sep 3, 2026monthly#6+7,539