See the nuclei.
Measure what matters.

BioNuclei is an automated analysis pipeline designed specifically for nuclear instance segmentation and quantification in fluorescence microscopy. AI predicts semantic regions; deterministic scientific software extracts individual nuclei, measurements and provenance.

Explore BioNuclei Open the Lab

ND2 / TIFFBoundary U-NetInspect results

The scientific workflow

One focused problem. One reproducible pipeline.

BioNuclei stays intentionally narrow: segment individual nuclei in fluorescence microscopy images, convert predictions into explicit instances, quantify them, and preserve the evidence needed to inspect the run.

BioNuclei workflow

Model, inference, deterministic instance extraction, measurements and provenance.

Explore the workflow →02

Analyses

Nuclear instances and quantitative measurements from supported fluorescence images.

View analyses →03

Formats

Supported Nikon ND2 and TIFF/TIF inputs with explicit acquisition handling.

View formats →04

Results

Segmentation masks, overlays, measurements and machine-readable provenance.

View results →05

Research

Training reference, evaluation, limitations and reproducibility records.

Read the research →06

BioMCP

Interoperability layer for exposing validated scientific operations to compatible AI agents and MCP clients.

Explore interoperability →

Separate sub-projects

A small ecosystem, with clear boundaries.

BioNuclei is the scientific core. BioMCP provides interoperability. ImageJ.JS offers manual browser based inspection in the Lab. The Lab is the user facing workspace that brings these routes together.

Core

BioNuclei

Automated nuclear instance segmentation and quantification.

Interoperability

BioMCP

Protocol boundary for compatible agents and MCP clients.

Manual

ImageJ.JS

Separate browser route for manual inspection.

Workspace

The Lab

Preview, inspect and continue into authenticated analysis.

Scientific software ecosystem

Trust and provenance

Evidence stays visible.

Validation records, reproducibility documentation and source code are part of the product surface rather than hidden implementation details.

Validation

Research record

Review staged validation, reproducibility status and open publication gates.

Results

Verified results

Inspect retained benchmark values and evidence context.

MIT License

Open source

Inspect implementation, documentation and provenance artifacts.

Scientific boundary

AI predicts. Deterministic scientific software measures.

BioNuclei is not a clinical diagnostic tool, not an online continuous-learning system, and not a general-purpose whole-cell imaging suite. Every result should remain inspectable and traceable to its computational provenance.