Chatcrystal by Zengliangyi is an MCP server and a local-first AI Personal Knowledge Management application designed for developers. It specializes in integrating with various AI coding tools to manage and distill coding conversations into an organized knowledge base. Developed primarily in TypeScript, it supports deployment across multiple environments including desktop, CLI/web, and cloud. Chatcrystal helps developers import, structure, and semantically search their AI-assisted problem-solving sessions, enhancing knowledge retention and retrieval for future projects. This free program is categorized as an mcps tool.

What does Chatcrystal connect to?

Chatcrystal connects to a variety of popular AI coding tools and services. These include Claude Code, Cursor, Codex CLI, Trae, and GitHub Copilot. Its primary function is to import and manage coding conversations originating from these platforms. The program acts as an MCP server, integrating with compatible AI clients such as Claude (Desktop and Code), Cursor, VS Code (Copilot), and Codex. This allows developers to centralize their AI interactions into a unified system for knowledge management.

Core features and tools

Chatcrystal provides several callable tools via its command-line interface. Key commands include crystal status for checking system state, crystal import [--source claude-code] for bringing in conversations, and crystal search "query" [--limit 10] for finding information. Developers can manage notes with crystal notes list [--tag X] and crystal notes get <id>, or summarize data using crystal summarize --all. Configuration is accessible via crystal config get, and the server can be started or stopped with crystal serve -d and crystal serve stop, respectively.

Who is it for?

Chatcrystal is intended for developers who regularly use AI coding tools to assist with problem-solving and code generation. The application supports users of platforms like Claude Code, Cursor, and GitHub Copilot, helping them organize and retrieve their AI-driven interactions. It serves those who wish to build a personal knowledge base from their coding conversations, turning ephemeral chat logs into structured, searchable notes. This is not the right fit if you do not use AI coding assistants to generate or refine code.

How to get started

The server authenticates using an API key passed as an environment variable. Specifically, clients configure integration using CHATCRYSTAL_API_TOKEN and communicate via STDIO.

Use cases

• Importing AI coding conversations directly from local tool data directories.
• Distilling raw conversations into structured notes, complete with titles, summaries, conclusions, code snippets, and relevant tags.
• Semantically searching the accumulated knowledge base, utilizing embeddings and relation-aware expansion for improved result accuracy.

Chatcrystal offers developers a distinct way to manage their AI coding conversations, transforming them into a searchable personal knowledge base. Its multi-platform distribution and integration with common AI tools make it a versatile utility for knowledge retention. By centralizing AI interactions, developers can efficiently revisit past solutions and insights. The use of environment variables for API tokens ensures client integration. Try importing conversations from Claude Code using the "crystal import" command.