AI-assisted 3D bioimage analysis platform. Replaces ImageJ/IMARIS with a modern browser-based interface -- no coding required.
Why CozyBio? Every bioimage lab we surveyed manually traces nuclei, counts cells by hand, and collapses 3D data to 2D because existing tools can't handle volumetric analysis without expensive software or custom code. CozyBio gives wet-lab biologists a free, AI-assisted pipeline: load a confocal z-stack, segment in 3D, measure per-object, review and correct in an interactive viewer, export results. No coding, no IMARIS license.
Status: Early development (v0.3). Core pipeline works: load data, 3D segmentation (Cellpose), morphometric measurement, OME-Zarr visualization. See CHANGELOG.md.
make setup # Install all dependencies
make dev # Start backend (:8000) + frontend (:5173)Open http://localhost:5173. Test dataset: Fluo-N3DH-CE on Zenodo.
make dev # Run both servers concurrently
make test # Backend tests (pytest)
make lint # Lint backend (ruff) + frontend (eslint)
make typecheck # Type-check frontend (tsc)
make format # Auto-format backend (ruff)
make check # lint + typecheck + test (run before PRs)backend/ FastAPI + Python (uv) — REST API, pipeline runner, OME-Zarr serving
frontend/ React + TypeScript (Vite + pnpm) — browser UI with neuroglancer 3D viewer
nodes/ Processing node definitions (Python) — segmentation, measurement, loaders
docs/ Architecture, PRD, feature history, feedback
See CONTRIBUTING.md for setup, guidelines, and how to add nodes. By submitting a PR, you agree to the CLA.
PolyForm Noncommercial 1.0.0 -- free to use, modify, and share for noncommercial purposes. For commercial licensing, contact nicelir1996@gmail.com.