Basalt is the platform to build and operate AI features : Craft high-quality prompts with our AI-powered Copilot, test and evaluate LLM outputs, deploy seamlessly with our SDK, monitor and refine performance in real conditions—all in a collaborative workflow.
Are you building an AI feature in your product? Tell me: How do you keep track of the prompts that are working and the ones that don't? How do you compare which models work better for you?
Until now I feel a lot of us would answer: a google sheet, or a notion page shared with the team 🤦♂️ But if you are serious about adding AI to your product, you can't just leave it to manual work.
That's why @Basalt is the perfect tool for engineering and product teams you get to iterate on the wording of each prompt, you get test your prompts as the models eveolve. At the end it gives you a dashboard to see performances. not anecdotal feedback.
The same way you would not write software without proper testing, you should not build AI feature without proper tooling.
I'm a user, love it! Congrats on building this highly valuable tool. I love the way you improve the imported prompts.
Integrating AI features in seconds is a bold promise, and I’m genuinely impressed!
Congrats on launching Basalt!
Just downloaded for the first time. Thanks for putting together! Looking forward to exploring.
Man, is Basalt useful.
I'm used to training, maintaining and deploying Machine Learning models in production, and trust me, you need a strong engineering team to go from prototype to useful model.
I was worried it would be a similar pain with LLMs: great prototypes, unusable in prod.
And here comes Basalt, which makes training, comparing models, evaluating performance over time, so easy an intern can do it.
I am about to release my new AI-native product - and ofc I used Basalt to : - Craft prompts using best-practices - Ensure the quality is good with 100+ test cases (without the grunt work, using custom & LLM-as-a-judge evaluators) - Figure out which models / parameters are best - Deploy in 30 sc in my codebase using the SDK - Monitor logs once in staging / production
It's such a great toolkit to build AI features!!
I haven’t tried Basalt yet, but I can’t wait to! It looks like a game-changer for AI feature development—finally, an all-in-one tool to build, test, deploy, and monitor LLM-powered features. Excited to see it in action!
I’m amazed ! This is the future of prompt engineering. Can’t wait to use it
The collaborative prompt testing workflow looks helpful for our team. Monitoring LLM performance in real conditions would solve some of our current deployment issues.
