Motif banner
motif-Labs motif-Labs

Motif

AI community

Description

A unification layer for AI coding agent sessions — collect Claude Code, Codex and Cursor sessions across your team, search them, hand one session natively to another tool, and let your agents recall what the team already knows. Self-hosted.

Installation

This entry records only its repository, not the path inside it, so there is no exact command to give. Open the source below and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

README

Motif

**A unification layer for AI coding agent sessions.** Open source, fully self-hosted.

Handing a Claude Code session off to Codex, natively

Start in Claude Code, finish in Codex — the session comes with you, natively.

Your team writes code with Claude Code, Codex, and Cursor — each tool keeps its sessions in its own format, in its own corner. Motif collects them in one place, makes them searchable across tools and teammates, streams them live to a team dashboard, and can hand a session started in one tool over to another tool **natively** — the target tool treats it as its own history, not a summary.

Features (v1)

  • Collect — a lightweight daemon watches Claude Code, Codex, and Cursor sessions on each dev machine and syncs them live to your self-hosted server. Sessions never leave your infrastructure. (Other open-source agent CLIs are on the roadmap; open-weight models themselves — Hermes, Qwen, Llama via Ollama/OpenRouter — already work today as the memory engine through the openai-compatible provider.)
  • Native handoff — convert a Claude Code session into a real Codex rollout file. codex resume picks it up as its own session; continue exactly where you left off.
  • Recall — an MCP server that hands your agents the 1-2k tokens that matter instead of letting them re-derive the codebase every session. Deterministic (FTS + session graph + human notes), so it needs no API key.
  • Ask a session — resume any past Claude Code or Codex session read-only on the machine that owns it and get an answer from the agent that lived it, not a summary.
  • Session memory — the server distills sessions into entity-based notes (files, decisions, topics) with supersession and conflict detection, powered by a pluggable LLM provider (Anthropic, OpenAI, any OpenAI