Gemini MCP Server

Instructions

Implementation Reference

• src/enhanced-stdio-server.ts:698-736 (handler)

  The handler function that executes the count_tokens tool. It takes text and optional model, calls the Gemini API's countTokens method, and returns the token count in MCP response format.

  private async countTokens(id: any, args: any): Promise<MCPResponse> {
    try {
      const model = args.model || 'gemini-2.5-flash';
      
      const result = await this.genAI.models.countTokens({
        model,
        contents: [{
          parts: [{
            text: args.text
          }],
          role: 'user'
        }]
      });
  
      return {
        jsonrpc: '2.0',
        id,
        result: {
          content: [{
            type: 'text',
            text: `Token count: ${result.totalTokens}`
          }],
          metadata: {
            tokenCount: result.totalTokens,
            model
          }
        }
      };
    } catch (error) {
      return {
        jsonrpc: '2.0',
        id,
        error: {
          code: -32603,
          message: error instanceof Error ? error.message : 'Internal error'
        }
      };
    }
  }

• src/enhanced-stdio-server.ts:308-323 (schema)

  Input schema definition for the count_tokens tool, specifying required 'text' parameter and optional 'model'.

  inputSchema: {
    type: 'object',
    properties: {
      text: {
        type: 'string',
        description: 'Text to count tokens for'
      },
      model: {
        type: 'string',
        description: 'Model to use for token counting',
        enum: Object.keys(GEMINI_MODELS),
        default: 'gemini-2.5-flash'
      }
    },
    required: ['text']
  }

• src/enhanced-stdio-server.ts:305-324 (registration)

  Tool registration in the available tools list, including name, description, and input schema.

  {
    name: 'count_tokens',
    description: 'Count tokens for a given text with a specific model',
    inputSchema: {
      type: 'object',
      properties: {
        text: {
          type: 'string',
          description: 'Text to count tokens for'
        },
        model: {
          type: 'string',
          description: 'Model to use for token counting',
          enum: Object.keys(GEMINI_MODELS),
          default: 'gemini-2.5-flash'
        }
      },
      required: ['text']
    }
  },

• src/enhanced-stdio-server.ts:480-481 (registration)

  Dispatch case in handleToolCall method that routes count_tokens calls to the handler function.

  case 'count_tokens':
    return await this.countTokens(request.id, args);