连接

WEFT

FPGArtkticbyFPGArtktic
•
开发者工具
学习与文档

Provides a safe, structured interface for LLM clients to interact with Intel Quartus Prime FPGA development tools.

WEFT (WEFT Elaborates FPGA Toolchains) is an MCP (Model Context Protocol) server designed to bridge large language models (LLMs) with complex FPGA development workflows, specifically targeting Intel Quartus Prime 25.1. It enables LLM clients to safely and efficiently perform tasks such as linting, simulation, asynchronous compilation, and report analysis, returning results as concise JSON instead of verbose log files. The tool containerizes execution of many tasks, ensuring a secure environment by validating all paths against a defined workspace root and operating with no network access at runtime. WEFT streamlines FPGA development by empowering AI models to interact with the toolchain programmatically, handling the plumbing so developers can focus on design.

主要功能

01Structured parsing of Quartus reports into concise JSON for resources, timing, and ranked messages.
02Integrated document RAG (Retrieval Augmented Generation) with OCR for contextual search and clause-level citations in FPGA documentation.
03Asynchronous management of Intel Quartus Prime 25.1 compilation projects with status and log retrieval.
04Secure, containerized execution of FPGA linting and simulation tools (Verilator, GHDL, Icarus) with waveform capture.
05Comprehensive source code indexing for project hierarchy, module information, and code search.
060 GitHub stars

使用案例

01Providing LLMs with contextual information from project source code and technical documentation for design assistance.
02Enabling LLMs to debug FPGA designs and testbenches by running lint and simulation commands programmatically.
03Automating FPGA compilation and analysis, providing structured feedback to AI assistants or automated workflows.