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 learns lessons from your sessions. Agent memory + context engineering as a single static Go binary.

Quick Install

Copy the config for your editor. Some servers may need additional setup — check the README.

Add to claude_desktop_config.json:

{
  "mcpServers": {
    "culi": {
      "command": "go",
      "args": [
        "run",
        "github.com/hung12ct/culi@latest"
      ]
    }
  }
}

README Excerpt

<div align="center"> <img src="docs/logo.png" alt="culi" width="120" height="120" /> <h1>culi ☕</h1> <p><b>Local context control for Claude Code and OpenAI Codex.</b></p> <p>One trusted knowledge store. Only the relevant cards reach each prompt. Useful lessons come back for review.</p> </div> **culi** (cà phê culi — the single dense peaberry bean) is the memory and context layer between

Tools (3)

anthropicollamaopenai

Topics

agent-memoryanthropicclaudeclaude-codeclaude-code-memoryclaude-memorycodexcontext-engineeringdeveloper-toolsembeddingsgolangharness-engineeringllmllm-memorymcp

culi

Quick Install

README Excerpt

Tools (3)

Topics

Related AI/LLM Servers