Back to browse
GitHub Repository

Local-first AI memory archive. Import ChatGPT, Claude, and Grok exports, generate semantic embeddings, and search via MCP server. Zero cloud, zero cost.

45 starsPython

MyChatArchive – bring your full ChatGPT history into Claude via MCP

by chan1·Mar 6, 2026·2 points·0 comments

AI Analysis

●●●BangerSolve My ProblemShip ItDark Horse

Local semantic search over your full AI conversation history, MCP-integrated, zero cloud.

Strengths
  • Solves genuine migration pain: ChatGPT's official export loses context; this preserves full transcripts and makes them searchable by meaning.
  • Auto-discovers Claude Code and Cursor sessions from local machine — no manual export required for many users.
  • True local-first: SQLite + sentence-transformers, zero API keys for core pipeline, respects privacy for sensitive conversation data.
Weaknesses
  • Requires Python 3.10+ setup and manual git clone—higher friction than a native installer or one-click executable.
  • MCP integration is solid but depends on Claude Desktop/Cursor adoption; unclear how search quality compares to ChatGPT's native retrieval.
Target Audience

Users migrating between AI assistants (ChatGPT to Claude, etc.) and developers using Claude Desktop, Cursor, or Claude Code

Similar To

Copilot's @workspace memory (context scope, not export) · Anthropic's MCP ecosystem (mechanism, not domain)

Post Description

Switched from ChatGPT to Claude and realized the official migration only transfers what ChatGPT remembers about you, not your actual conversations. Built a local pipeline that imports full exports, generates semantic embeddings on your machine, and serves them via MCP server. Claude Desktop and Cursor can search your entire chat history by meaning during any conversation. No cloud, no API keys for core pipeline. Also supports Claude Code, Cursor, and Grok exports.

Similar Projects

AI/ML●●Solid

OpenContext – Bring Your Own Coding Agent, Local-First, No Vendor Lock

The MCP server to give Claude true persistent memory is the clearest win — it tackles the annoying, real-world problem of lost context across sessions. The repo pairs local analysis (Ollama) with import/export for ChatGPT/Claude/Gemini and ships as Node/TypeScript + Docker with a usable web UI, so you can run everything without vendor APIs; I'd like to see wider model/import support and clearer UX around merges/conflicts, but the core idea and execution feel immediately useful.

Niche GemSolve My Problem
akarnam37
103mo ago