Self-improving context & memory for Claude Code and OpenAI Codex. One canonical knowledge store injected only when relevant — hooks push budgeted context, an MCP server serves depth on demand, and it
culi ☕ Local context control for Claude Code and OpenAI Codex. One trusted knowledge store. Only the relevant cards reach each prompt. Useful lessons come back for review.
culi (cà phê culi — the single dense peaberry bean) is the memory and context layer between your coding agents and your projects. It stores rules, lessons, styles, patterns, skills, and agent definitions as small Markdown cards, then routes a token-budgeted selection to Claude Code or Codex.
It is local-first, git-backed, and inspectable. Hooks and retrieval make no model calls; optional background learning proposes new knowledge for review. No Node, Python, or resident daemon is required.
- One cross-agent memory for Claude Code and Codex, with explicit claude: and codex: provenance. - Small, relevant context through hooks, with deeper on-demand retrieval through MCP. - Background learning through Codex CLI, OpenAI API, Claude CLI, Anthropic API, or Ollama. - Codex history recovery from its read-only local rollout database, automatically throttled and deduplicated when hooks do not provide a transcript path. - A production-style local console for review, knowledge maintenance, delivery activity, Knowledge Pulse analytics, backend configuration, and light/dark themes. - Plain files as truth under ~/.culi/knowledge/; SQLite holds rebuildable card-search tables plus preserved local activity state.
Without culi With culi ------ Guidance is copied across CLAUDE.md, AGENTS.md, agents, and skills One canonical, git-versioned card store Every session loads large instruction files Only prompt-relevant cards a
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/hung12ct--culi.html)
Your agents already solved this. deja finds it — retroactive search and MCP recall over the session logs your coding agents already wrote. Cross-agent, SSH-portable, one zero-dep binary, fully local.
cuongnbms/mastering-claude-codeCẩm nang thực chiến Mastering Claude Code (tiếng Việt): 16 chương biên soạn từ tài liệu chính thức của Anthropic — CLI, workflows, MCP, subagents, skills, hooks, plugins, CI/CD & tối ưu context.
replenhq/replenSmarter AI Development Workflows
rwnalds/engramSelf-hosted MCP server + dashboard giving Claude Code, Cursor & Hermes agents shared long-term memory over a git-backed markdown vault. Agent-native, Obsidian-compatible second brain — markdown RAG, n
Smitner-Studio/facetRender agent-written Markdown as a live local page, and send the human's click back to the agent as typed data. Human-in-the-loop approval gates for coding agents.
0xwilliamortiz/openclaude-improvedruns anywhere. uses anything
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.