Picsellia is the all-in-one platform to manage, deploy, and monitor computer vision models at scale. From dataset versioning to experiment tracking and real-time inference monitoring, it’s built for ML teams working with visual data. Streamline your CV workflow and stay in control.
Picsellia Atlas
Hey Product Hunt! 👋
Honestly, we're so excited (and maybe a little nervous!) to finally share Atlas with you all today.
As AI engineers ourselves, we kept hitting the same wall: general AI models are impressive, but quite overkill and shitty for custom VisionAI use-cases. Supervised fine-tuning is still king.
But in the same time LLMs are becoming so good, so we though what if? What if LLMs could help us with data-centric AI?
Humans, paired with AI agents, can curate Datasets, evaluate model trainings, and monitor models waaaay faster.
So, we built Atlas – an open-source AI agent designed to tackle this head-on.
Think of it as the Data-centric expert co-pilot we wished for. It helps you:
Analyze datasets (finding labeling issues, imbalances, etc. – the stuff that trips up models!)
Suggest the right model architectures for your specific task
Train, evaluate, and iterate much faster
Here's why we think it's different, especially for this community:
Speed & Precision: Seriously, our goal is to help you go from data to robust custom models way quicker.
It's Open Source: Because we believe in transparency. No black boxes – inspect it, fork it, help us make it better! We want this to be a community effort.
LLMs never sees your Data: This was non-negotiable for us. Atlas uses abstract representations and never sends your actual images externally. Your competitive advantage stays safe.
Atlas can integrate with platforms like Picsellia (where we initially built it!) or run standalone (coming soon). Honestly, we built Atlas for people like us: ML engineers, developers, researchers tired of reinventing the wheel for every new vision project.
We've poured our hearts (and countless hours!) into making powerful VisionAI development more accessible and transparent for everyone.
We'd be thrilled if you check out the repo, give Atlas a spin, and let us know what you think! Seriously, all feedback is gold ; tell us the good, the bad, and the buggy! 🙏 https://github.com/picselliahq/atlas
Ask us anything below, the team and I will be hanging out here all day (fueled by coffee ☕️)!
Happy Building!
Thibaut & the rest of picsellia Team (@ Picsellia)
@thibaut_lucas Amazing work!!
@thibaut_lucas @henrik_balling
Congratulations on the launch, Atlas makes vision workflows feel lightweight and modern great work!
Picsellia Atlas
Hello folks 👋
Pierre-Nicolas, Picsellia CTO here 🫡
I'm soooo happy that we finally get to release this 🔥
We've been in the Computer Vision field for a while now and "AI Agents" has been written in our roadmap since the first days of GenAI hype.
But it took us some time to find how we wanted to do it, we didn't want to add just a "GPT-like" feature in the product like everyone else.
But now I can tell you that we found the perfect recipe so YOU will get the most value out of Picsellia AND AI Agents 😎
I'll be there all day to answer your questions and maybe provide you support right on Picsellia 😉
Thanks for your support 🙌
Tate-A-Tate
🎉 Congrats on the Atlas launch! As an product lead, I’m thrilled to see a tool that tackles the real pain points of VisionAI workflows. Love how you’ve fused LLMs with data-centric AI while keeping data private 🔒—huge win for teams wary of leaks! The open-source angle is 🔥 for community trust.
Quick suggestion: Adding pre-built templates for common use cases (e.g., object detection in retail) could help new users ramp up faster. Any plans for this?
Can’t wait to try the standalone version! 👏
Picsellia Atlas
@rocsheh Thanks for your warm words 🙏
I'm glad you really understood the philisophy of the product!
Indeed, the first analysis we built are kind of templates for common Object Detection and Segmentation tasks, no matter the use-cases. But by leveraging more tools built precisely for specific industries and also the power of the agenst that will be able to decide which tool to call and when, this is precisely what we will be able to offer 🎉
Also as it is open source it should be pretty easy for anyone to add tools and analysis that are more suited for their use-case and just do a drop-in replacement, that's how we intend the repository to go ☺️
Can't wait to release the standalone version and contribution guidelines to see how people build with it 💪