# janreges/ai-distiller-mcp [Health: Active]

**Category:** 🔒 Security  
**Repository:** https://github.com/janreges/ai-distiller  
**GitHub Stars:** 167  
**npm Downloads (last month):** 200  
**Views:** 3  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/janreges-ai-distiller-mcp

## Description
Extracts essential code structure from large codebases into AI-digestible format, helping AI agents write code that correctly uses existing APIs on the first attempt.

## Tools
Capabilities this server exposes over MCP:

- **distill_file** — Extract structure from a single file
- **distill_directory** — Extract structure from entire directories
- **list_files** — Browse directories with file statistics
- **get_capabilities** — Get info about AI Distiller capabilities
- **aid_hunt_bugs** — Generate bug-hunting prompts with distilled code
- **aid_suggest_refactoring** — Create refactoring analysis prompts
- **aid_generate_diagram** — Produce diagram generation prompts (Mermaid)
- **aid_analyze_security** — Generate security audit prompts (OWASP Top 10)
- **aid_generate_docs** — Create documentation generation prompts
- **aid_deep_file_analysis** — Systematic file-by-file analysis workflow
- **aid_multi_file_docs** — Multi-file documentation workflow
- **aid_complex_analysis** — Enterprise-grade analysis prompts
- **aid_performance_analysis** — Performance optimization prompts
- **aid_best_practices** — Code quality and best practices prompts
- **aid_analyze** — Direct access to all AI actions for custom workflows

## Claude Desktop Quick Installation
Install path inferred — verify against the README before running it. Uses `npx` (confidence: medium):

```json
"mcpServers": {
  "ai-distiller-mcp": {
    "command": "npx",
    "args": ["-y","@janreges/ai-distiller-mcp"]
  }
}
```

## Documentation

## What janreges/ai-distiller-mcp MCP server does

janreges/ai-distiller-mcp MCP server makes the AI Distiller code-structure extractor available to MCP-compatible clients. It is intended for projects where sending the entire source tree to an AI model would exceed practical context limits or make API relationships difficult to identify.

The core operation reduces source code to a compact representation of the information an agent usually needs to use existing code: public methods, properties, types, imports, and basic documentation. The default output excludes method bodies and protected, internal, or private members. Filtering options can include those members, implementation details, comments, or other content when a broader view is required.

The repository describes support for more than 12 programming languages and provides text, Markdown, structured JSON, JSONL, and XML output formats. Distillation can target one file or a directory, and the tool can report compression and processing statistics.

## How it works

The MCP interface separates browsing, extraction, and analysis-prompt generation. `list_files` helps an agent inspect directories and view file statistics. `distill_file` and `distill_directory` then produce reduced code context for a file or directory. `get_capabilities` reports available AI Distiller capabilities.

The analysis tools do not perform the underlying review themselves. Instead, they prepare prompts containing distilled code for an AI agent to process. Available workflows cover bug hunting, refactoring, Mermaid diagram generation, OWASP Top 10 security analysis, documentation, systematic file-by-file analysis, multi-file documentation, complex analysis, performance, and coding best practices. `aid_analyze` provides direct access to the AI actions for custom workflows.

## Setup and configuration

Install the MCP package published as `@janreges/ai-distiller-mcp` from npm, then configure it in an MCP-compatible client. The README explicitly names Claude, Cursor, VS Code, and Windsurf as supported integration targets, and includes a Claude Desktop usage guide in its documentation structure.

No API key or external paid AI service is specified in the provided material. The generated prompts are saved to a `.aid/` directory by default for workflows that create files; small codebases can instead use standard output according to the CLI behavior described by the project.

Configuration follows AI Distiller's filtering model. Public members, documentation, and imports are included by default, while implementation bodies and non-public visibility levels are excluded. Worker count, output format, comments, docstrings, and summary style can be adjusted through the underlying Distiller options.

## Tools and capabilities

The MCP tool set includes:

- File and directory distillation
- Directory browsing with file statistics
- Capability discovery
- Bug-hunting and refactoring prompt generation
- Mermaid diagram prompt generation
- OWASP Top 10 security-analysis prompts
- Documentation and multi-file documentation workflows
- Deep file analysis and complex analysis workflows
- Performance and best-practice analysis prompts
- A general analysis entry point for custom actions

## Limitations and notes

AI Distiller prepares context and prompts; it does not carry out the requested security review, refactoring, documentation task, or other analysis itself. An AI agent must read or execute the generated prompt content.

The project identifies this MCP release as an early version and requests feedback through GitHub discussions or issues. The supplied material does not state a license, so licensing should be verified in the repository before redistribution.

_Full upstream README: https://allmcps.com/mcp/janreges-ai-distiller-mcp/readme_