OKF Agent Memory OKF Agent Memory
A Git-native persistent memory layer for AI coding agents that stores knowledge as structured Markdown files within the repository. It uses in-memory BM25 search and progressive disclosure to provide fast, auditable, and context-efficient memory without external database dependencies.
The project released version 0.2.0 in September 2026, introducing code_refs for binding concepts to source paths and a --for-path search flag to improve retrieval accuracy based on file location.
As of
- Sub-300µs in-memory BM25 keyword search
- Embedded Model Context Protocol (MCP) server
- Progressive disclosure for optimized token usage
- Trust tiers distinguishing human-verified vs. agent-generated content
- Zero-dependency Go binary with native Git integration
Long-lived repositories with multiple agents, or engineering teams requiring auditable, human-reviewed architectural decision records that persist across agent sessions.
Projects requiring semantic (vector-based) search across synonyms, cross-repository memory sharing, or small, simple projects where maintaining a structured knowledge bundle adds unnecessary overhead.
Free and open-source under the MIT license, with no external database, API, or embedding costs.
An excellent, performant alternative to opaque vector databases for teams that value auditability, Git-native workflows, and minimizing infrastructure complexity.
Is OKF Agent Memory free?
Yes - OKF Agent Memory is Open Source. Free and open-source under the MIT license, with no external database, API, or embedding costs.
What is OKF Agent Memory best for?
Long-lived repositories with multiple agents, or engineering teams requiring auditable, human-reviewed architectural decision records that persist across agent sessions.
Who makes OKF Agent Memory?
OKF Agent Memory is developed by okf-memory. It is listed in the AI Codebase Knowledge & Generation category on ai.dosa.dev.
Favorite this tool to revisit it later, or Zap it to contribute to the public vote count.
Content on this page is AI-generated. Please verify details with the vendor's website for accuracy.