Basalt - Integrate AI in your product in seconds

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.

Add a comment

Replies

Best

The Basalt Copilot is a lifesaver! It helps refine prompts intelligently, suggesting better phrasing and structures to get the best outputs. No more trial and error! I've been building outbound sequences using Basalt connecting it directly to my Data studio thanks to the SDK!


Go team!

 The amount of people telling us the time they save with our Copilot is insane! Gogogo Quentin 🔥

Hello AI builders ! Super happy to be a part of this amazing team. Evaluating AI features is tough, but we live and breath by trying to solve this pain point.


What makes Basalt different from other LLM ops tools out there : we consider this to be a team job between product, data and tech teams, instead of a simple dev tool. We think this is the best way our customers can reach for state of the art LLM integrations.


We're launching in San Francisco, get in touch !

Francois and the Basalt team

 So proud to work on this with you 💪 Let's go!

Awesome! Love that you can quickly compare multiple model outputs.

 It's definitely a killer feature Maxime 💪

  If you see any model provider we should add please ask us !

Definitely a must-have solution for any company serious about building AI products 🫡

 Thank you for your support Nicolas 🙌

Love the testing-first approach! 🚀 Curious how Basalt handles team collaboration on prompt versions? Seems like a game-changer for AI dev workflows!

 thanks Robert for the positive comment. Indeed collaboration building great AI features is key across the tech team involving product, eng, data, domain experts that evaluate quality.

Collaboration is pretty simple for the moment in the tool :
- we built a UI that makes Basalt a welcoming place for non-tech people
- having prompts out of the code base with our SDK-get-prompt feature is key to ensure every-one can test the feature
- versioning, annotation ... everything is built in the plateform with user logic ensure good collab on key gestures !

I hope it helps, if you see missing and useful collab features we would love to hear from you !

Best,

Vic

 Thank you Robert! Indeed we handle collaboration on prompting and evaluation :)

Let's go Basalt team!

 Thank you for your support Thomas! 🙌

 Thank you Thomas !

This is such a much-needed tool for many product builders out there! The prompt iteration struggle is real, so many teams rely on intuition rather than structured testing, which leads to inconsistent AI performance. Love that Basalt streamlines everything from crafting to deploying AI features with real-world evaluation.


Congrats on the launch!

Best wishes and sending lots of wins to the team :)

  Thanks so much Shivam on the support, you nailed our positioning perfectly well !

We are building the second generation of LLM Ops tools, aiming at helping teams achieve feature quality rather than "just" technical monitoring

 Thanks for your message Shivam :)

A product for Software Development ease with proper tools that monitor performance with automated tracking and insights that too with Deploying any changes on the go which makes it better at development journey.

 Thank you Ajay! Can't wait to see what you build on Basalt 🙌

 You nailed it about the ability to avoid bottleneck with developers deploying yet another prompt template !

Let's go Basalt team!! I'm a fan of the platform already :)

I rate the focus on high-quality prompts w/ the AI-powered Copilot, definitely going to be optimising my interactions & outputs from language models.


I feel like your approach is allowing for the perfect streamlined & collaborative workflow :)

Best of luck today & look forward to following the journey!!

 Thanks for your support Sam! We really want to make Basalt the best framework ever (AI powered) to help everyone create high quality features faster, as a team 🤘

 Thanks so much for the feedback Sam !

I love the LLM as a judge option. Congrats on the great work !

Thanks Etienne! It's one of my favorite features too! We've taken the time to refine our evaluation models to drastically reduce the number of false positives/negatives and give the user the ability to give only what the AI needs to evaluate, without being an assistance engineer. 🪨