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)
Keep the signal. Cut the noise. Local CLI + MCP — fold verbose tool output for Claude Code, Codex, Cursor & Desktop.
magna-nz/tallybookPrices your Claude Code and Codex sessions from the transcripts on disk, shows what each agent cost over any period, and tells you in plain English what would have been cheaper. CLI, Browser UI and MC
ParsaVictor/code-context-engineLocal-first MCP context engine for AI coding agents — cuts token cost ~90% via a per-project code graph + code folding. Works with Cursor, Claude Code, Codex & any MCP client. Web + ML/PyTorch codebas
azimov777/casefileThe task tracker your AI agents keep for each other: every task carries a case file, so the next agent picks up where the last one stopped. MCP server + live board. Self-hosted, one-line install.
okf-memory/okf-agent-memoryGit-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zer
vshulcz/deja-vuYour 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.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.