greptile-mcp

Instructions

Implementation Reference

• src/server.ts:358-370 (handler)

  The handler function that executes the greptile_help tool. It returns a text content block containing comprehensive help documentation retrieved from getHelpContent().
  /** * Handle greptile_help tool */ private async handleGreptileHelp(): Promise<{ content: Array<{ type: string; text: string }> }> { return { content: [ { type: 'text', text: await this.getHelpContent(), }, ], }; }
• src/server.ts:86-93 (schema)

  Input schema definition for the greptile_help tool (empty object as it takes no parameters), registered in the ListTools response.
  { name: 'greptile_help', description: 'Get comprehensive help and usage examples for all Greptile MCP tools', inputSchema: { type: 'object', properties: {}, }, },
• src/server.ts:225-226 (registration)

  Registration and dispatch of the greptile_help handler in the CallToolRequestSchema switch statement.
  case 'greptile_help': return await this.handleGreptileHelp();
• src/server.ts:584-714 (helper)

  Helper function that generates the comprehensive markdown help content used by the greptile_help tool.
  private async getHelpContent(): Promise<string> { return `# 🚀 Greptile MCP Server - Comprehensive Guide ## Overview The Greptile MCP Server provides AI-powered code search and querying capabilities through the Model Context Protocol (MCP). It integrates with the Greptile API to index repositories and answer natural language questions about codebases. ## Available Tools ### 1. \`greptile_help\` Get this comprehensive help documentation. **Usage:** No parameters required ### 2. \`index_repository\` Index a repository to make it searchable for future queries. **Parameters:** - \`remote\`: "github" or "gitlab" - \`repository\`: Repository in "owner/repo" format - \`branch\`: Branch to index - \`reload\`: Force reprocessing (optional, default: true) - \`notify\`: Email notification (optional, default: false) **Example:** \`\`\`json { "remote": "github", "repository": "microsoft/vscode", "branch": "main", "reload": true } \`\`\` ### 3. \`query_repository\` Query repositories using natural language to get detailed answers with code references. **Parameters:** - \`query\`: Natural language question - \`repositories\`: Array of repository objects (optional if session exists) - \`session_id\`: For conversation continuity (auto-generated if not provided) - \`stream\`: Enable streaming response (optional, default: false) - \`genius\`: Enhanced capabilities (optional, default: true) - \`previous_messages\`: Conversation history (optional) **Example:** \`\`\`json { "query": "How is authentication implemented in this codebase?", "repositories": [ { "remote": "github", "repository": "microsoft/vscode", "branch": "main" } ], "stream": false } \`\`\` ### 4. \`get_repository_info\` Get information about an indexed repository. **Parameters:** - \`remote\`: Repository host - \`repository\`: Repository identifier - \`branch\`: Branch name ## Best Practices ### Session Management - Use consistent \`session_id\` across related queries for better context - Auto-generated session IDs are provided if not specified - Session continuity improves answer quality over multiple interactions ### Repository Workflow 1. **Index** repositories first using \`index_repository\` 2. **Verify** indexing status with \`get_repository_info\` 3. **Query** using natural language with \`query_repository\` ### Query Optimization - Be specific in your questions for better results - Use technical terms relevant to the codebase - Reference specific features, patterns, or components - Build on previous queries using session continuity ## Example Workflows ### Getting Started with a New Codebase \`\`\` 1. Index the repository 2. Ask: "What is the overall architecture of this codebase?" 3. Follow up: "How are the main components organized?" 4. Dive deeper: "How does the authentication system work?" \`\`\` ### Debugging a Specific Issue \`\`\` 1. Ask: "How is error handling implemented?" 2. Query: "What are the common error patterns in this codebase?" 3. Follow up: "Show me specific examples of error handling code" \`\`\` ### Understanding Design Patterns \`\`\` 1. Query: "What design patterns are used in this codebase?" 2. Follow up: "How is the observer pattern implemented?" 3. Query: "Show me specific implementations of this pattern" \`\`\` ## Configuration Required environment variables: - \`GREPTILE_API_KEY\`: Your Greptile API key - \`GITHUB_TOKEN\`: GitHub personal access token Optional: - \`GREPTILE_BASE_URL\`: Custom API base URL (default: https://api.greptile.com/v2) ## Resources - **greptile://help**: This documentation - **greptile://config**: Current server configuration ## Prompts - **codebase_exploration**: Guided codebase exploration with focus areas --- 🔗 **More Information:** Visit the [Greptile documentation](https://docs.greptile.com) for detailed API reference and advanced usage patterns.`; }
• src/index.ts:73-137 (handler)

  Alternative inline handler implementation for greptile_help in the Smithery-compatible index.ts module.
  // Register greptile_help tool server.registerTool( 'greptile_help', { title: 'Greptile Help', description: 'Get comprehensive help and usage examples for all Greptile MCP tools', inputSchema: {}, }, async () => { const helpContent = `# 🚀 Greptile MCP Server - Comprehensive Guide ## Overview The Greptile MCP Server provides AI-powered code search and querying capabilities through the Model Context Protocol (MCP). It integrates with the Greptile API to index repositories and answer natural language questions about codebases. ## Available Tools ### 1. \`greptile_help\` Get this comprehensive help documentation. ### 2. \`index_repository\` Index a repository to make it searchable for future queries. **Parameters:** - \`remote\`: "github" or "gitlab" - \`repository\`: Repository in "owner/repo" format - \`branch\`: Branch to index - \`reload\`: Force reprocessing (optional, default: true) - \`notify\`: Email notification (optional, default: false) ### 3. \`query_repository\` Query repositories using natural language to get detailed answers with code references. **Parameters:** - \`query\`: Natural language question - \`repositories\`: Array of repository objects - \`session_id\`: For conversation continuity (auto-generated if not provided) - \`stream\`: Enable streaming response (optional, default: false) - \`genius\`: Enhanced capabilities (optional, default: true) ### 4. \`get_repository_info\` Get information about an indexed repository. **Parameters:** - \`remote\`: Repository host - \`repository\`: Repository identifier - \`branch\`: Branch name ## Best Practices ### Repository Workflow 1. **Index** repositories first using \`index_repository\` 2. **Verify** indexing status with \`get_repository_info\` 3. **Query** using natural language with \`query_repository\` ### Session Management - Use consistent \`session_id\` across related queries for better context - Session continuity improves answer quality over multiple interactions For more information, visit: https://docs.greptile.com`; return { content: [{ type: 'text', text: helpContent }], }; } );