Research & summarize
Crystallize Code Repositories into Searchable Knowledge
Transforms code repositories into AI-digestible knowledge bases, enabling efficient search and understanding of complex projects with a 5:1 compression ratio.
Without it
Stand up the this tool server yourself — bot tokens or OAuth, a process to host and keep alive, updates to chase.
With it
Install once and it is wired straight into your agent.
What you get
- Initialize repository crystallization and scan files.
- Generate comprehensive analysis guidance and quality standards.
- Create searchable knowledge bases with natural language queries.
- Monitor crystallization progress and validate context quality.
Connect it yourself
Use hosted
Use this with MCP Gate
Connect this and other MCP tools to your agents through one endpoint, with usage billed in one place.
An AI Context Engineering tool that transforms large repositories into crystallized, AI-understandable knowledge through systematic analysis and optimization, creating searchable knowledge bases for AI agents.
Installation
NPM Global
npm install -g context-crystallizer
NPX
npx context-crystallizer
Configuration
Claude Desktop
{
"mcpServers": {
"context-crystallizer": {
"command": "npx",
"args": ["context-crystallizer"],
"cwd": "/path/to/your/project"
}
}
}
Available Tools
| Tool | Description |
|---|---|
get_crystallization_guidance |
Get comprehensive analysis guidance with templates and quality standards |
init_crystallization |
Initialize repository crystallization, scan and prepare files |
get_next_file_to_crystallize |
Get the next file for AI analysis in the crystallization process |
store_crystallized_context |
Save generated AI knowledge and crystallized context |
get_crystallization_progress |
Monitor crystallization status and progress |
search_crystallized_contexts |
Find relevant knowledge by functionality or keywords |
get_crystallized_bundle |
Combine multiple contexts within token limits |
find_related_crystallized_contexts |
Discover code relationships and dependencies |
search_by_complexity |
Find contexts by complexity level (low, medium, high) |
validate_crystallization_quality |
Assess context quality with completeness and readability metrics |
update_crystallized_contexts |
Update contexts for modified files and maintain accuracy |
Features
- Functionality search with natural language queries
- 5:1 compression ratio (source code to crystallized context)
- AI-optimized format and structured for LLM consumption
- Smart bundling that combines multiple contexts within token limits
- Automatic .gitignore compliance and binary file filtering
- Searchable crystallized knowledge base creation
- Progress monitoring and quality validation
- Code file relationship detection
- Context filtering by complexity
Usage Examples
I need to understand how authentication works in this massive project
What files depend on the authentication system?
How does authentication work in this app?
Show me how the payment system works
What depends on this Auth.tsx file?
Notes
Inspired by AI Distiller, focused on enabling AI agents to generate comprehensive crystallized contexts about functionality, patterns, and relationships. The server should run from the project root directory for integration with AI agents. Crystallized contexts are saved in the .context-crystallizer/ directory.