What's New

July 13, 2026
Click to learn more
Culture Monitor

Accurate Well Image Proportions

Well and montage images now preserve the microscope's native aspect ratio instead of being forced into a square, so round wells render as circles rather than stretched ovals.

Culture Monitor

Custom Metadata Columns on the Cultures Page

Add and display custom metadata as columns on the Cultures Page table, aggregated across each plate's cultures.

Culture Monitor

Smarter Assistant Chat Scrolling

AI Assistant chat now anchors on send and follows new content only when you're already at the latest message, with a pill to jump back to unseen replies.

Culture Monitor

Bulk-Set Metadata from the Cultures Page

Set cell line, cell line lot, passage number, and custom key/value metadata across many cultures or wells at once from the Cultures Page.

June 30, 2026
Click to learn more
On-Prem Agent

Metadata Presets for Keyence Ingestor

Save and reuse plate barcode and magnification presets for Keyence (BZ) ingestion, with auto-generated dataset folder names.

June 30, 2026
Click to learn more
On-Prem Agent

Yokogawa CQ3000 Microscope Support

Ingest data from the Yokogawa CQ3000 high-content microscope, with montage details in dataset summaries.

June 21, 2026
Click to learn more
Culture Monitor

AI Assistant Approves Culture Status Changes in a Single Step

Asking the assistant to update multiple cultures at once now requires just one approval, instead of hanging mid-task after the first few confirmations.

June 21, 2026
Click to learn more
Culture Monitor

AI Assistant Report Generation from Picklist CSV

Attach a plate-well picking-list CSV to the AI Assistant to generate a targeted report covering only the selected wells.

June 21, 2026
Click to learn more
Culture Monitor

Culture Activity History Timeline

Track every status change, metadata edit, and observation on a culture in a newest-first activity timeline. (early access)

June 21, 2026
Click to learn more
Culture Monitor

Classification Well Confidence Filtering

Filter classification wells by model confidence — adjust a per-label 0–100% slider to surface only the predictions you trust.