Lauterbach TRACE32® AI Extension

This extension brings TRACE32® AI to Visual Studio Code. Chat with the TRACE32® AI Assistant and let AI agents control PowerView through the TRACE32® MCP server.

For more information, see Getting Started with TRACE32® AI.

TRACE32® AI Assistant

The TRACE32® AI Assistant is a retrieval-augmented generation (RAG) system based on the official documentation, primarily designed to answer TRACE32®-related questions. It provides semantic search in the TRACE32® manuals, PRACTICE syntax checks and a product search in the Lauterbach product catalog. The TRACE32® AI Extension makes it available as a chat participant and a MCP server.

Talk to the AI Assistant from the Chat window by prepending your message with @t32chat. Before your first message, the extension logs in at the t32chat backend and stores the device token securely for future sessions.

The assistant is additionally registered as a remote MCP server (t32chat). To start the MCP server, use the command MCP: List Servers.

TRACE32® MCP Server

To use the TRACE32® MCP Server, t32mcp and TRACE32® PowerView must run on the same system, with the Remote API enabled.

The TRACE32® MCP Server (t32mcp) gives AI agents access to TRACE32® PowerView.

• Download the recent release from the GitLab release page
• Configure its path with the setting lauterbach.ai.t32mcp.path

The lauterbach.ai.t32mcp.connection setting selects how the extension connects to t32mcp:

• stdio (default): Visual Studio Code launches the t32mcp executable as a child process and communicates over stdio
• Full HTTP URI (e.g. http://127.0.0.1:3000): the extension connects to a t32mcp instance you have already started with --http. In this mode the path, policy, skills, and rcl settings are not used; you pass them on the t32mcp command line yourself. If that instance was started with --token, set the token setting to have the extension send it as Authorization: Bearer header.

To start the MCP server, use the command MCP: List Servers.