Wavedrom-LLM Instructions
Table of Contents
Introduction
Available Commands
Privacy
Introduction
Wavedrom-LLM is a tool to to produce beautifully rendered timing diagrams straight from your text prompts. It builds upon the original Wavedrom project (by the way one of my favorite tools to generate timing diagrams) and takes it one step further from having to code JSON directly. Wavedrom-LLM builds on the Google's Gemini API, and with system instructions can understand and construct wavedrom-compliant JSON. With multi-turn instructions (keeping track of the history) you can incrementally change diagrams. One additional trick, you can also directly edit the JSON to update timing diagram, and on the next LLM construction, it will update the model with your changes. If you want to add your comments, suggestions, or improvements, use the github project below, which is a fork on the original Wavedrom project
Wavedrom-LLM project on GitHub - Issues tracking + discussion
Wavedrom classic website
Available Commands
Here are types of commands that can be used: Generate a clock named PCK with 10 negative edges Add a signal PDAT with 4 cycles logic zero followed by 6 xs Stretch all signals by 4 cycles Break all the signals at the 4th cycle Group together the PCLK and PDAT signals with the group name Inputs Add the data ADDR0, RW, and DATA in the SDA signal with yellow state Clear the old signals and generate an I2C waveform Add a space after SCL
Privacy
So by nature, your instructions and timing diagrams are running through our server, and then through Google Gemini API. Please don't share anything proprietary! As with every AI-based solution, your examples, use cases, types of diagrams, and corrections are stored and can be used to help improve the model. If you need a truely private option, ping me. info@wavedrom-llm.com