OpenViking
by volcengine
OpenViking is a database that stores what an AI agent knows (documents, memories and skills) as a browsable virtual filesystem. It helps agents find relevant information faster and lets people read and edit what the agent has stored.
- Stars
- 41k
- Licence
- AGPL-3.0
- Runs on
- A small server
- Setup
- Technical
- Language
- Python
- Age
- 296 days
What you can do with it
- Give an AI agent long-term memory that you can open, read and edit as Markdown files
- Load project documents, code and web pages so an agent can search them by folder
- Store reusable skills and session history in one place for agents to retrieve
Make money with it
A freelancer could set up and maintain agent memory and knowledge-base systems for small businesses as a paid service.
An idea, not financial advice. Check the licence before charging for anything built on it.
Ways a business could use it
- Add persistent memory to a customer support or sales assistant so it remembers client preferences and past conversations
- Build an internal knowledge base that an agent can search, cutting time spent answering repeat questions
Can you use it commercially?
Free to use. Anything you distribute that includes it must be open-sourced under the same licence. Running it as an online service also counts, so a hosted version must offer its source.
Not legal advice. Read the project's licence file before building a product on it.
Getting it running
You set up a server using uv and Python 3.10 or later, and connect a model provider that supplies an embedding model and a vision-language model.
Follow the official instructions in the project README.
Things to check
- AGPL-3.0 licence requires sharing source changes if you offer it as a network service
- Needs a model provider for embeddings and a vision-language model, which may mean paid API keys
- Benchmarks were run with Volcengine models, so results may differ with other providers
Who's talking about it
Checked 28 Oct 2026. Summary drafted by AI from the project's README.
