Overview
Turn scanned PDFs and faxes into editable documents, and pasted lab reports into checked values with units, reference ranges and H/L flags. The original stays alongside, and the clinician confirms before anything is saved.
How it works
PDFs are read page by page with OCR. Scans are cut into high-resolution tiles so numbers, units and drug names survive the request-size limit, and transcription runs at temperature 0 to avoid ‘plausible’ corrections. The text becomes an editable draft the clinician corrects before saving, with an optional cleanup of scan artifacts that never touches medical content: numbers, units and dosages come back character-identical. Pasted lab text goes to an agent that returns every value as typed JSON (parameter, value, unit, reference range, collection date). The values are reconciled with the practice’s parameter catalog (unit conversion, reference ranges, H/L flags) and applied as a new report after the clinician ticks each value.
What Corti handles
- Reading scanned pages into text
- Transcription that keeps what is written, errors included
- Pulling typed lab values out of free text
What you build
- Upload and paste points in your product
- Your parameter catalog and the matching rules
- The review screens and saving to your record
Make it yours
- Write the instructions the agent follows
- Pick a larger model for harder problems, or a smaller one for speed and cost
- Pick the model the agent runs on
- Skip the thinking step when you want faster, cheaper answers
- Connect the tools and data sources it can use
- Choose how much effort the model puts into reasoning
- Decide who can see and use the agent
- …and much more