AssayPilot Bench · DishFlow
Point. Shoot.
Review the weird ones.
Take one plate photo or upload the whole batch. AssayPilot proposes colony-like objects, coverage and size summaries on the server, then gives you an overlay you can correct instead of counting from zero.
No app store required. The current detector runs on AssayPilot infrastructure; this release does not send your plate images to an external AI API.
The 30-second workflow
Batch first. Exceptions second.
1. Photograph plates without opening them just for the photo.
2. Select the whole batch and optionally paste plate IDs.
3. Review orange/dense/low-quality results first.
4. Use auto suggestions as the starting point for human marks.
5. Export both layers with provenance.
What it measures now
Useful image measurements, not biological theatre.
Colony-like candidate count, candidate coverage, size distribution, plate localization, blur and glare warnings, repeated-plate time series.
It does not identify species, infer viability, calculate CFU/mL without dilution/plated-volume metadata, or decide TNTC for every protocol.