# volcengine/OpenViking

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

Repository: https://github.com/volcengine/OpenViking
Canonical: https://ross.abutalabs.com/products/openviking
Homepage: https://openviking.ai/
Language: Python
License: AGPL-3.0
License Family: copyleft
Topics: context-database, agentic-rag, agent-memory, self-evolving, agent-plugins, dsh-plugin
Last push: 2026-08-26T16:22:58+00:00

## Health v2 (maintenance only)
Score: 78/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 17
- inputs: {"age_days": 240, "days_push": 7, "days_rel": 9, "gap_med": 2, "n_releases_24m": 74}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 33569, forks 2553 (observed 2026-08-28T04:11:58.334832+00:00)

## What it is
OpenViking is an open-source context database for AI agents that unifies agent memory, knowledge RAG, and skills into a single virtual filesystem under the viking:// protocol. Content is tiered into abstract, overview, and detail levels and loaded on demand, with retrievable, debuggable retrieval trajectories.

## Use cases
- give my AI agent persistent memory across sessions
- store and retrieve knowledge for an agent without a black-box vector store
- let an agent browse its own context like a filesystem
- combine agent memory, RAG knowledge, and skills in one store
- debug what context my agent retrieved and why
- build an agentic RAG pipeline with on-demand context loading

## When to choose
- you're building AI agents that need unified memory, knowledge, and skill context
- you want filesystem-style navigation (ls, tree, find) over agent context instead of raw vector queries
- you need tiered context loading to reduce token usage
- you want inspectable retrieval trajectories for debugging agent behavior

## When to avoid
- you need a general-purpose production vector database for non-agent workloads
- you require a permissive license — OpenViking is AGPL-3.0
- you need a fully managed hosted service rather than self-hosted infrastructure
- your project only needs simple key-value memory without retrieval semantics

## Facets
- artifact type: library
- maturity: active
- function: vector-database, rag, agent-framework, search-engine, file-system
- domain: artificial-intelligence, large-language-models, databases
- platform: python, self-hosted, cross-platform
- tags: context-database, agent-memory, agentic-rag, virtual-filesystem, skills, self-evolving, llm-context, ai-agents, retrieval-augmented-generation, docker

## Member repositories
- volcengine/OpenViking (main) score 78

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:58.334832+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:28:18.267405+00:00, confidence not recorded.
  - readme: https://github.com/volcengine/OpenViking (fetched 2026-08-28T04:11:58.334832+00:00, sha ea4e2fcaf68f)
  - homepage: https://openviking.ai/ (fetched 2026-08-29T07:48:36.105665+00:00, sha 111129be4d9f)
  - registry_pypi: https://pypi.org/pypi/openviking/json (fetched 2026-08-29T07:48:36.108229+00:00, sha 3105bda98bea)
- Data as of 2026-08-30T08:39:29.467469+00:00.