#codex-cli

MCP server bridging AI assistants to OpenAI Codex CLI for code analysis and review

Stop re-explaining your codebase to AI every session. CogniLayer gives Claude Code persistent memory across sessions, AST-based code intelligence (who calls what, blast radius analysis), and subagent context compression (40K → 500 tokens). 17 MCP tools including semantic search (FTS5 + vector), code_context, code_impact, session bridges, crash recovery, and deployment safety gates. Saves 80-200K+ tokens per session by replacing file reads with targeted knowledge retrieval. Works with Claude Code and Codex CLI. Zero-config after install — hooks and MCP tools run automatically.

An MCP server that scales development into controllable agentic, recursive flows, and build a feature from bottom-up