Skip to main content
← All projects
Dafe Ofoh / Senior DevOps & AI Engineer02 / Clinical AI / documentation

Fanoni Scribe

An ambient documentation workflow that keeps a draft SOAP note connected to its source transcript for clinician review.

Scribe product design
Fanoni Scribe visit review screen showing a sample transcript beside a SOAP note with source-line citations
Captured from the Scribe repository’s interactive design prototype. Sofia Ramirez and the visit are sample data. View full-size image ↗
In practice

From conversation to reviewed note

A clinician records a visit after consent, then checks the draft against the transcript before signing. This screen shows the side-by-side review view, with note citations highlighting source lines.

  1. CaptureRecord the visit and preserve a source transcript.
  2. DraftOrganize the conversation into SOAP sections with source citations.
  3. ReviewLet the clinician check and edit the draft before signing.

The issue

A generated note can omit context or introduce unsupported detail unless the clinician can compare it with what was said.

How it works

The workflow records a visit, drafts a structured SOAP note, connects note content to transcript evidence, and supports review before FHIR R4 output. Its repository describes a stack including GCP Cloud Run, Vertex AI, Postgres, and Redis/BullMQ.

Discuss a project