Basalt
Reach 99% quality on your AI feature
Start new thread

Basalt Agents - Evaluate AI workflows and reach 99% AI quality.

by•
Basalt, #1 AI observability tool for teams, is launching its brand new Agent Builder : prototype, test, and deploy complex AI flows composed of multiple prompts, and run them through a dataset of scenarios.

Add a comment

Replies

Best
Maker
📌
Thanks Nicolas for the hunt ! I’m François, co-founder of Basalt, and together with the team we’re super excited to share our brand new feature: the Agent Builder 🚀 What pain are we solving ? We're experts in AI observability, and along the way we noticed something important: engineers and PMs want to prototype AI agents before shipping them, but workflow builders are not made for evaluations. What’s new with our Agent Builder Basalt's Agent Builder is the easiest way to design, test, and optimize complex AI flows built from sequences of prompts and actions. 👉 Create and customise your own flows 👉 Chain prompts and actions with ease 👉 Run each prompt on the best-suited model 👉 Set custom evaluation criteria for each step 👉 Measure performance on hundred of real case scenarios with dedicated datasets Stop testing prompts in isolation. Turn your flow into a single, end-to-end workflow you can evaluate thoroughly, and then deploy with confidence. Hope you enjoy it ! ❤️ from the Basalt team
Maker

So proud to launch that with you 💪

 Congrats on the launch team!

Basalt is back on Product Hunt, 6 months after an epic fight!! Let's go, love the product!!

Maker

 thanks for the love !!

Maker

 Still in war mode!

Hunter
Friends don’t let friends building AI apps without proper logging, and evaluation. Basalt is the go to tool, easy to install, simple to setup. Great to see the improvements from v1
Maker

 thanks so much for the Hunt !!!

Maker

Thank you for your support !!!

Maker

Love the energy on the video ! At Hexa we are already addicted to your agent builder ! Can't wait to see what's coming next !

Maker
Maker

Good to be back on Product Hunt! 🚀

I’m Guillaume, co-founder of Basalt, and today we’re introducing a major step forward in how teams build with AI: the Basalt Agent 🎉

Until now, prototyping agents meant juggling between workflow tools, prompt sandboxes, and spreadsheets, but none of them were really made for evaluation.

That’s the gap we’re closing.

With Agent Builder, you can:

🔗 Chain prompts and actions effortlessly

🤓 Assign each step to the best model for the job

📐 Define your own evaluation metrics

📊 Run large-scale tests on real-world scenarios

From your first chained prompt, to complex multi-turn agentic workflows, Basalt gives you a single place to design, test, and optimize, before you ever ship.

This one’s for everyone who’s ever said, “I just want to see how it performs in context.”

Now you can.

Thank again to for the hunt! I hope you’ll enjoy using it as much as we enjoyed building it 💛

— The Basalt team

Maker

Excellent! Can agents hit my own retrieval system when testing?

Maker

 Yes it can! You can use API Request node to ping your system or even soon a code block :)

Maker

 absolutely ! You can implement your own retrieval system in the workflow builder and test how well it performs

That’s a solid upgrade! Congrats.

Maker

 Thanks !!

Guys ! so happy to have you back on Product Hunt, the platform has powered up so much congrats !

Maker

 thanks Vic !!

Basalt Agents is an impressive tool—being able to prototype, test, and optimize complex AI workflows can really boost team productivity. Do you plan to add real-time monitoring or performance analytics for these AI agents?

Great launch! How does Basalt prevent overfitting on the evaluation dataset — e.g. if prompts start optimizing too much to test cases and lose generality?

Maker

 great question !! To avoid overfitting, we recommend creating a dynamic dataset, meaning that you continuously enrich it with new test cases from your logs (something you can do from Basalt or programmatically) !

12
Next
Last