Raycast Local File Search with LLM Integration

This Raycast extension provides powerful local file search capabilities with content indexing and LLM-powered analysis features.

Features

Search Capabilities

• Fast local file search with content indexing
• Fuzzy matching for filenames and content
• Snippet preview with highlighted search terms
• File type filtering
• Detailed metadata display

LLM Integration

• File summarization using Anthropic's Claude AI
• Ask questions about file content
• Get AI-powered insights about your files

Setup

Important First Steps

1. Clone or fork the repository
2. Install dependencies: npm install
3. Configure the folder to be indexed:
   • Open src/services/mcp/index.ts
   • Locate the constructor method (around line 52)
   • Change the default path from Downloads to your preferred directory:

   // Change this line
   this.downloadsPath = path.join(os.homedir(), 'Downloads');
   
   // To something like this for your Documents folder
   this.downloadsPath = path.join(os.homedir(), 'Documents');
   
   // Or for multiple directories (advanced)
   // You'll need to modify the listFiles method as well
   this.downloadsPath = path.join(os.homedir(), 'YourPreferredFolder');

4. Create a .env file in the root directory with your Anthropic API key:

   ANTHROPIC_API_KEY=your_api_key_here
   

Starting the Application (Two-Step Process)

1. First, start the server:

   npm run dev:server
   

   This will start the backend server that handles file indexing and search.

2. Then, in a separate terminal window, start Raycast:

   npm run dev
   

   This will start the Raycast extension.

3. The extension should now be available in Raycast

Usage

Basic Search

1. Open Raycast and select "Search local files"
2. Type your search query
3. Browse through the results
4. Use the file type filter to narrow down results

LLM Features

• Select a file from the search results to see AI-generated context about the file
• Ask natural language questions about your files
• Get AI-powered insights based on file content

Advanced Usage

• For more precise searches, use specific keywords
• For natural language queries, phrase your question clearly
• The LLM will analyze the most relevant sections of your files based on your query

Troubleshooting

• If search results aren't appearing, make sure the server is running (npm run dev:server)
• If the LLM features aren't working, check that your API key is correctly set in the .env file
• If you change the indexed directories, restart the server for changes to take effect
• If you're getting errors about file permissions, make sure the directories you're trying to index are accessible

Development

Prerequisites

• Node.js
• Raycast
• Anthropic API key

Project Structure

• src/search-local-files.tsx: Main UI component
• src/llm-utils.ts: Utilities for LLM integration
• src/extension.ts: Extension lifecycle management
• src/server/index.ts: Backend server for file indexing and search
• src/services/mcp/index.ts: File system operations and content extraction
• src/services/search/index.ts: Search indexing and query processing

License

MIT