rekal
Gives LLMs long-term memory backed by hybrid search in a portable SQLite file, so agents remember and retrieve relevant context reliably.
MCP server that gives AI coding agents long-term memory across sessions, stored locally in a single SQLite file (~/.rekal/memory.db). Retrieval blends BM25 keyword search, 384-dimension vector embeddings, and recency decay, all computed locally with no cloud, daemon, or API keys. Works with Claude Code, Codex CLI, and OpenCode; the companion Claude Code plugin adds hooks that steer the agent to store and recall memories.
- Store durable preferences and decisions in one session and have Claude recall them in later sessions
- Search project memory by meaning, keywords, and recency when resuming work
- Share one local memory store across Claude Code, Codex CLI, and OpenCode
Gives LLMs long-term memory backed by hybrid search in a portable SQLite file, so agents remember and retrieve relevant context reliably. Simple to deploy and move, with no external database required.
Developers building agents that need durable, searchable memory without the overhead of managing a separate database service.
https://github.com/janbjorge/rekal
By janbjorge
How to Get It
uv tool install rekal && claude mcp add --scope user rekal -- rekal && claude plugin marketplace add janbjorge/rekal && claude plugin install rekal-skills@rekal
Tip: Paste this into a Claude Code conversation. Verify command matches your Claude Code version.
After installing, paste this into Claude:
Help me store durable preferences and decisions in one session and have Claude recall them in later sessions
Trust Signals Auto-scanned
Community Pulse Emerging
Discussed on Hacker News
- Show HN: Rekal – Long-term memory for LLMs in a single SQLite file — Hacker News · 9 pts
1 mentions across 1 sources
Reviewer notes
Auto-scanned review. These are observations, not a security certification.
catalog_hygiene stale-eval refresh: Scored from trust signals (evidence-eval-v1): 52 GitHub stars; 1 contributors; last commit 16d ago; license MIT.
Things to check
- Requires disabling Claude Code's built-in auto memory (add "autoMemoryEnabled": false to ~/.claude/settings.json) or the two systems compete; the plugin's hooks only partially mitigate this. Early-stage project with limited community feedback so far.
- Single maintainer. Consider the risk if this person stops maintaining the project.
How to evaluate tools before deploying →
Data shown here comes from public APIs and automated scanning. Reviewer notes reflect one person's experience. This is not a security certification or legal recommendation. Always evaluate tools according to your own organization's policies.
Evaluation
catalog_hygiene stale-eval refresh: Scored from trust signals (evidence-eval-v1): 52 GitHub stars; 1 contributors; last commit 16d ago; license MIT.