The problem: When Claude Code creates an implementation plan, it lives in your terminal session. Nobody else sees it until it becomes a PR. If you want GPT to check the architecture or a teammate to flag issues, you're copy-pasting between windows.

This MCP server fixes that. When your agent creates a plan, it gets shared as a collaborative thread in CoChat. Engineers comment on it, other AI models review it, and you pull all the feedback back into your agent's context with one command. Decisions can be saved as project memories that persist across sessions and are searchable by anyone.

What it does:

Plans: Auto-shared as collaborative threads. Pull feedback back into your terminal. Cross-model review: Have GPT review your Claude plan, or vice versa. Project memories: Semantic memory that persists across sessions, models, and people. Ask: Query your project's knowledge base from the terminal. Auto-scoping: Detects your project from git remote. No config needed. Setup is one command per agent. Auto-share behavior is configurable (off/plan/all).

MIT licensed, available on npm: npx @cochatai/mcp-cochat

Happy to answer questions about the architecture or the MCP protocol integration.