Software built for your science.

Generic software wasn't built for your science. BioDevs develops software fitted to your experimental workflow, to increase scientific throughput, reduce human error, and measure parameters generic tools cannot.

Discuss your needs See our work

It starts with a discussion, or a visit to your lab, to identify what is worth building for your experimental workflow.

The gap

Your science is specific. Your software isn't.

ImageJ, Excel, GraphPad and R are everywhere in the lab, and none was designed for your scientific questions. Manual analysis with them limits your scientific throughput, is prone to human bias and error, and cannot measure the biologically relevant parameters your question depends on.

Custom software for

generating standardized protocols

Standardized, custom protocol generation – dilution calculations, setup diagrams – to eliminate avoidable human error.

controlling your instruments

Software to drive your cameras, microscopes and test rigs, fitted to what your setup actually needs.

automating repetitive tasks

From simple chores like data copy-paste to advanced data-science pipelines.

analyzing images & video with AI

Deep-learning image and video analysis, with AI models trained specifically for your task.

cleaning & structuring your data

Automated cleaning and structuring of your raw data, ready for analysis.

publication-ready figures & statistics

Analysis, visualization and publication-ready figures, with the right statistical tests built in.

Our mission

Software built for your experimental workflow

We build what your work needs: from simple automation of computer-based tasks (protocol generation, copy-pasting), to controlling your scientific instruments, to analyzing your data with advanced data science and training AI models for complex image or video analysis.

Why BioDevs

We understand biological relevance and software science

We work at the interface of experimental biology and software. That combination turns raw data acquisition into biologically meaningful measurements: less manual work, fewer errors, parameters generic software cannot extract, and orders of magnitude higher scientific throughput.

Higher throughput Fewer human errors New biological parameters Publication-ready output Higher-impact publications

Beyond efficiency, custom software reduces observer bias and improves the reproducibility and reliability of your results, making them easier to defend in review and helping your work reach higher-impact journals. The software itself can also be published.

Built by Théo Mercé, PhD

PhD in biochemistry and toxicology, developer and data scientist specialized in deep-learning biological analysis, author of the DanioTracker, ZOTAS and DanioAChE tools.

Case study

Zebrafish Research

Deep-learning applications and lab software delivered for toxicology and drug discovery on zebrafish larvae. Proof that highly specific scientific needs become robust tools.

High-throughput acquisition of escape responsesTracks every larva, even when they crossAutomatic detection of seizure-like events

DanioTracker

Custom video tracking built on vision AI models trained by our team: analysis over 100× faster than with generic tools, without expert intervention. It extracts domain-specific parameters for toxicology and pharmacology (neuromuscular evaluation, postural control, epileptiform seizure detection), fully automated from raw video to statistics and publication-ready figures.

Explore the platform

ZOTAS

Deep-learning quantification of adiposity in the Zebrafish Obesogenic Test (ZOT) from fluorescence images, for obesogen screening and risk assessment: 50× faster image analysis, without manual work. With it, the team identified chemicals that induce an obesogenic thrifty phenotype, independent of physical activity and food intake.

Published in:

Biology Methods and Protocols Communications Biology · Nature Portfolio

DanioAChE

Image-analysis software measuring acetylcholinesterase (AChE) inhibition in larvae, for nerve-agent and pesticide toxicology and antidote screening. Custom-trained AI extracts brain and muscle regions to quantify histochemical features and toxicological or pharmacological potency.

Explore the platform

ProtocolGen

Web app that turns a list of compounds into a ready-to-use experimental protocol: automatic dilution calculations, multi-well plate layouts and exposure schedules, exported as a standardized PDF to share with colleagues and add to calendars.

ZebraDash

Dashboard that centralizes the parameters extracted by DanioTracker across all experiments: select the compounds and endpoints to compare, get replicates grouped automatically, and generate publication-ready figures in one click: dose-response heatmaps, radar profiles and raincloud plots.

Contact

Let's discuss your needs

Tell us about your experiments and scientific problems. We'll identify the bottlenecks and where custom software would help.