autospectra

AutoSpectra is an all-in-one MCP server providing browser automation, API testing, debugging, and integration with AI agents via the MCP protocol. It exposes tools for navigating browsers, performing clicks and text input, sending HTTP/GraphQL requests, validating responses against schemas, and running interactive debug sessions. Developers can orchestrate these tools through the MCP protocol using the use_mcp_tool helper, enabling workflows that combine browser automation with API validation and debugging in a single automated pipeline. The server is designed to work with Claude, ChatGPT, Gemini, Perplexity, and other MCP-enabled agents, and supports both visible and headless browser modes for debugging and automation efficiency. To get started, configure your environment, start the server, and invoke tools like navigate, click, api_request, validate_schema, and debug_test via the MCP interface.

Prerequisites:

• Node.js installed (recommended LTS version)
• Git available to clone the repository
• optional: Docker if you prefer containerized runs

Step-by-step installation:

1. Clone the repository: git clone https://github.com/your-username/autospectra-mcp-server.git cd autospectra-mcp-server

2. Install dependencies: npm install

3. Build the project (if a build step exists in this repo): npm run build

4. Create an environment file if required:

   • Create a .env file in the repository root and set variables like PORT, DEBUG, ANTHROPIC_API_KEY, HEADLESS, SLOW_MO, OUTPUT_DIR, etc.

5. Start the server:

   • In development or production, use the MCP interface (e.g., npm start or the configured node command) depending on how you expose the entry point. Example: npm start

6. Optional Docker usage:

   • Build and run via npm scripts if provided (see Docker support section in the README): npm run docker:build npm run docker:run

Environment variables can customize behavior:

• PORT/DEBUG/HTTP_SERVER control core server behavior
• ANTHROPIC_API_KEY enables Claude integration features of the Claude computer use tools
• HEADLESS and SLOW_MO adjust browser automation for debugging and performance
• OUTPUT_DIR stores artifacts generated during runs Common issues include missing environment variables, incorrect API keys, or port conflicts. Ensure .env is properly configured and that the MCP client (AI agent) is configured to communicate with the server's MCP endpoint. When debugging, run in headful mode (HEADLESS=false) to visually inspect browser automation, and use the provided debug_test tool to step through test scripts. If using Docker, ensure Docker daemon is running and that the npm scripts align with your containerization approach.