Hi there! I’m César, founder of Latitude. I’ve talked to dozens of companies building AI features. If you’re not assessing the quality of your LLM outputs, you’re not serious about AI. Latitude is the open-source prompt engineering platform to build, evaluate, and refine your prompts with AI. This is how it works: 1. Create or paste your prompt into our Prompt Editor 2. Evaluate the output in batch - using an existing dataset or generating a synthetic one 3. Iterate your prompt with an AI-powered refiner If you increase the accuracy of your prompts, you’ll create more value for your users. It’s that simple. I’m actively looking for feedback. Let me know what you think!

Super cool product @heycesr . How do your evaluation prompts work? Can you define them yourself or are they done dynamically? Also, how about the prompts themselves, I assume you as the developers, have access to all of our prompts, correct?

What an amazing Friday! Love it! What an outstanding work César! Keep up the great work with Latitude

congrats on the launch! 🚀 I've seen prompt dev/eval platforms before but none this slick. Great job on the examples folder, so cool and fast to get up and running to do chain-of-thought, conditionals and JSON schemas.

Wow Cesar! It's super cool. Being able to build and refine prompts is key for every founder working on AI. Glad to see you helping on this and wish you all the best here!

Latitude user here! For us, a tool like this is absolutely essential. In the world of LLMs, you can achieve a 10/10 result multiple times, only to get a 1/10 suddenly, and then back to 8/10. It’s just too unpredictable. Having a way to manage this, iterate on prompts, and refine results is incredibly valuable. Best of luck with the launch from a satisfied user!

You can tell pros designed this product.

@heycesr I'm curious, how do you handle batch evaluations that run into the hundreds of test cases? Do you have built in throttling/delay when running in batch? We find that we typically run into rate limits when evaluating prompts on large datasets,

Congratulations guys on product launch 🚀. There is a issue with landing website - it's redirecting you to signup page when you click `star on Github` button.

Bravo on launching Latitude! The AI boom needs more tools like this to help developers harness the full potential of LLMs.

Looks amazing! I have been asking myself how to take advantage of LLMs in my products. The demo make this super clear, and the platform makes total sense. Keep up the great work :)

Congrats on the launch! 🚀 Latitude's open-source platform for prompt engineering is a great tool for developers looking to build and refine their LLM features with precision. Excited to see how it will enhance the development process and empower teams!

I recommend Latitude, an open-source platform for prompt development. Join and get an additional 10,000 launches to unlock the full potential of your projects!

🤩

Congrats on the launch César!!! Really curious and excited about this I have already created an account and started playing a bit - too soon yet to give feedback I think the product is quite useful but a bit hard to understand at the beginning. I gotta say the demo explains the product very clearly and in just 3 mins. I think that being able to use the latitude API to make the evaluations of my prompts and review them in the dashboard it's pretty awesome and something I'd probably pay for. Also, I think you could drive more attention to that feature in the demo (i know you gotta go straight to the point haha) With that said, this is a very useful tool and well packaged Congrats César and all the Latitude team Seguimos !!!

I feel like adding more documentation and guides would make it so much easier for beginners to get started.