Hi HN! This is Mateo, Maria, Faisal, and Osse from Countify (https://www.countifybio.com/). We're building a collaborative no-code computer vision platform for life sciences. Think Replit, but for scientific image analysis. Labs use us to build, train, and automate AI image analysis workflows and experimental design without writing a single line of code.

Almost every experiment in biology produces images: microscopy, cell cultures, tissue slides, well plates. Someone has to analyze them. Today that means one of two bad options.

Option one: a scientist counts cells and traces regions by hand. It's slow, it's subjective, and it's nearly impossible to reproduce. Option two: someone on the team learns to code — Python, ImageJ macros, deep learning models — which most biologists never will.

The tools that already exist don't close the gap and spread out throughout the workflow in capture, analysis, and data management. ImageJ and CellProfiler are powerful but have brutal learning curves. Arivis and others costs $30k+ per year and are for very specific use cases. Off-the-shelf AI models don't generalize to your specific assay and need to deal with tiny datasets that don't exist anywhere else. Building a custom model that actually works on your data takes an ML engineer, and academic labs and early-stage biotechs don't have one. So image analysis becomes the bottleneck that stalls the entire experiment.

Countify runs as a visual-intelligence agent that works alongside you in real time. Point it at your image; with any specimen, any experiment and it detects, reporting as data comes in. No code, and no manual annotation. You build and automate the workflow once; it runs the same way every time. Reproducible by default.

We serve a single kind of user that is in spread across industry and academia, small teams performing frontier research in biotech that don't have access to automation, however the platform scales organically to enterprise.

Countify is built around collaboration. Any user can publish the workflows they build to the whole network. Contributors get paid when others run their work. Consumers skip the build step entirely and get a workflow that's already been validated by someone who does that exact assay for a living. Everyone in the network compounds everyone else's work.

Where are we today?

We're early, and we're not going to dress it up. We have two recurring users: one academic lab and one CRO. Its small, but we began ramping GTM and we're adding roughly one new user a week, and onboarding a second CRO in August.

How we got here:

We met at Northeastern's incubator where we had one week to go from idea to MVP to pitching investors. We won the competition. But more importantly, we walked out with real interest from labs at Harvard, Northeastern, and Stanford; and quickly after 30+ waitlist signups from other labs as top research institutions. We knew we were solving something real and had to dive in head first.

We’d appreciate any feedback, questions, or advice. Thanks for reading.

— Mateo, Maria, Faisal, and Osse