Local ChatGPT-to-Markdown bridge for MCP-enabled AI workflows
chatgpt2md, from NextStat, migrates ChatGPT conversation history into a local, searchable Markdown archive that other models can reference. It converts exported vendor data into a machine-readable archive and exposes past discussions so alternative models can reuse prior context. Key functionality centers on export-to-Markdown conversion and making archives accessible to MCP-capable clients. The tool targets AI power users, developers, and researchers who need portable conversation records and a migration path away from the original service. It addresses vendor lock-in by making archives readable without the original service.
What tasks can you actually use it for?
The app accepts ChatGPT export packages produced through Settings > Data Controls > Export Data, including the ZIP/JSON bundle received by email, and turns those files into organized Markdown documents. It automatically structures saved conversations by year and month, producing a folder of plain-text files suitable for manual review, code-based processing, or importing into other local tools that read Markdown.
How reliable is its local search and retrieval?
The tool builds a local full-text index using the Tantivy library, which provides fast keyword and phrase lookup across converted conversations. It exposes MCP endpoints such as search_conversations, get_conversation, and list_conversations so host applications can query specific threads. That design produces near-instant retrieval on typical desktop hardware, improving access to past discussion points without scanning raw files manually.
How practical is deployment inside an MCP workflow?
chatgpt2md ships as a single binary with no external runtime dependencies, and the developer supplies pre-built executables for macOS (Intel and Apple Silicon), Windows, and Linux. The tool offers one-command setup for Claude Desktop and Claude Code, and requires an MCP-compatible host application to serve the indexed history to other models. The developer builds the project in Rust, which explains the compact single-file distribution.
What privacy and migration limits should you expect?
The tool processes exports entirely on the local machine, it does not upload conversation data to external servers, and the index and Markdown files remain on disk. Migration is limited to ChatGPT export formats, so moving histories from other vendors requires separate conversion. Community feedback on GitHub and Reddit notes the practical benefit of making years of context usable within MCP-enabled clients.
Who should use this tool and when
chatgpt2md is a focused migration utility for technically capable users who can obtain their ChatGPT export and run an MCP host, offering a dependable way to preserve machine-readable conversation history. It is not aimed at casual users who lack exported archives or an MCP client. Practical tip: clean and organize your export before indexing to improve later search relevance and retrieval precision.
