[GitHub Trending] volcengine/OpenViking

7 relevance

Self-evolving context database for AI agents directly addresses agent memory and RAG, highly relevant.

AI/ML github.com

Answer the board on AI readiness

Arrive with evidence, not opinion. See how your foundation benchmarks against your peers, get a clear read, and leave with a ranked place to start.

Why It Matters

For engineers building AI agents, this replaces opaque vector stores with a deterministic, observable filesystem-like context layer that cuts costs and improves accuracy, directly addressing the token waste and debugging pain points in multi-agent and long-session workflows.

Key Takeaways

Evaluate OpenViking as your agent's context store to reduce token consumption by up to 91% while boosting retrieval accuracy, using its tiered loading and observable retrieval paths for debugging.

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

Summary

OpenViking is an open-source context database for AI agents that exposes memories, resources, and skills as a virtual filesystem under the viking:// protocol, enabling agents to browse context with commands like ls and find instead of querying a black-box vector store. Content is processed into three tiers (L0 abstract, L1 overview, L2 details) and loaded on demand, reducing token spend, while each retrieval leaves a debuggable trajectory. Benchmarks show accuracy gains from 24-57% to 80-83% on long-conversation user memory (LoCoMo) with 34-91% fewer input tokens, and task success improvements of +6.87pp (retail) and +11.87pp (airline) on tau2-bench.

Author

volcengine