Loading video...
Video Failed to Load
curl This is our Meta-Prompt. It allows LLMs to understand our Reader, Embeddings, Reranker, and Classifier APIs for improved codegen. Using the meta-prompt is straightforward. Just copy the prompt into your preferred LLM interface like ChatGPT, Claude, or whatever works for you, add your instructions, and you're set. In... show more
209,078 views • 1 year ago •via X (Twitter)
10 Comments

and here's something cool - there's this handy CLI tool called llm from @simonw that lets you pipe the meta-prompt into an LLM right from your termail. So just do curl | llm -s 'TASK: grab all sentence from hackernews frontpage and embed and visualize the results in a 2d umap with matplotlib' -m claude-3-sonnet You'll get clean generated code output in your terminal. If you're using Cursor IDE, you can also add the meta prompt from to its docs index to help it understand our search foundation APIs better.

The meta-prompt is maintained under We'd love to hear your thoughts on how we can improve the meta-prompt. So give it a try and let us know how it works for you.

Are there any docs in how to create such a meta prompt? Looks amazing

@domasbitvinskas you need to do a metaprompt like this for @superinterfaces too

Smart. Many more should do this.

This is fantastic! I hope this becomes a trend. More and more API maintainers should be doing this (a win-win). Is the meta-prompt hand-crafted, or is there a more efficient way to generate/maintain it for a given API. Does having OpenAPI support help?

@weremeow new standards

Whats the prompt for generating the meta prompt? Super valuable resource for any set of technical docs.

Genius

Nice job
