Discover and explore top open-source AI tools and projects—updated daily.
JiaboLi-GitHubAI-driven GPU frame analysis and debugging
Top 98.0% on SourcePulse
Summary
renderdoc-mcp provides an AI-native interface for RenderDoc, enabling AI assistants to programmatically analyze GPU frame captures and debug graphics pipelines. It exposes RenderDoc's replay API through a structured MCP server and CLI, allowing automated inspection, debugging, and export of capture data without manual UI interaction, benefiting AI developers and researchers.
How It Works
The project wraps the RenderDoc replay API, transforming its functionality into 59 structured tools accessible via an MCP server and a command-line interface. This architecture allows AI assistants like Claude and Codex to open .rdc captures, navigate frames, inspect pipeline states, debug shaders and pixels, analyze resource usage, and export evidence programmatically, streamlining complex GPU debugging workflows.
Quick Start & Requirements
renderdoc-mcp.exe), CLI (renderdoc-cli.exe), and bundled RenderDoc runtime. Client configuration for AI tools like Codex Desktop is automated or can be manually added to ~/.codex/config.toml.cmake and the RenderDoc source code. Build command: cmake -B build -DRENDERDOC_DIR=<path-to-renderdoc-source> && cmake --build build --config Release.Highlighted Details
Maintenance & Community
No specific details regarding contributors, sponsorships, or community channels (e.g., Discord, Slack) were found in the provided README.
Licensing & Compatibility
renderdoc-mcp.Limitations & Caveats
The project requires access to the RenderDoc source code for building from source. No other explicit limitations, alpha status, or known issues were detailed in the provided README.
1 month ago
Inactive