A guided tour

Six features. One clinical workflow.

Scroll through how OmniscribeAI turns a visit into a signed, source-checked note — and what each step gives back to the clinician.

01 · Capture

The problem

Clinicians lose hours to documentation for every hour of care.

Document the visit while it happens.

Record live, upload a voice file, or paste text — all in one screen. Attach the patient first, and the draft is grounded in this encounter instead of a blank page.

  • One screen for live audio, files, and pasted text
  • Resume an interrupted recording without losing audio
  • The note is half-written before you sit down to chart
A microphone dissolving into sound waves that resolve into a written note
The Transcribe screen on a phone
02 · Trust the output

The problem

An AI note is only useful if the clinician can defend it.

AI drafts. You sign. Always grounded.

Every note is reviewed and locked with a signing PIN, and a source check flags any line not grounded in the visit before you ever sign it.

  • A defensible, attributable note — who drafted, who signed, when
  • Source check catches copy-forward errors before the chart
  • Copy the whole document or section by section into the EHR
A locked document card with a verification checkmark and source links
A signed clinical note on a phone
03 · Sound like your specialty

The problem

A generic SOAP note doesn't fit a PT eval or a therapy session.

97 templates, tuned to how you practice.

Specialty frameworks span Rehab, Medical, and Behavioral Health across 60+ specialties. Clone any built-in into your own and the AI follows that exact format every time.

  • Nearly 100 built-in frameworks, ready to use
  • Clone and customize — one house style for the whole team
  • Prose, bullet, or hybrid — choose before drafting
A fanned deck of layered note-template cards
The template library on a phone
04 · Remember every visit

The problem

Every visit shouldn't start from a blank chart.

The patient's story carries forward.

Prior signed notes, active diagnosis codes, and recent measures travel with the patient. Trends in pain, range of motion, or PHQ-9 surface at a glance before the next visit.

  • Carried diagnosis codes nudge you when today's note diverges
  • Objective measures trend up, down, or stable over time
  • Behavioral Health stays compartmentalized from other divisions
A flowing thread connecting glowing nodes along a timeline with trend graphs
The patient context screen on a phone
05 · Ask anything

The problem

Answers live in two places: the chart and the literature.

Ask your charts. Ask the research.

Miss Cleo answers questions from signed notes under BAA, then switches to web-backed medical research that never touches PHI — all without leaving the visit.

  • Signed-notes Q&A on the patient in front of you (BAA-covered)
  • Research mode with citations and a strict no-PHI guard
  • A floating panel that follows you across every screen
A glowing assistant orb beside a speech bubble and an open book
06 · Built for HIPAA

The problem

Medical AI has to earn trust before it earns adoption.

HIPAA-grade by construction. Pay for what you use.

BAA gates every PHI feature, each PHI access is logged, and the clinical and non-clinical AI run on entirely separate paths. Credits show exactly what each note and minute cost.

  • BAA gate, separate clinical/non-clinical AI, full PHI audit log
  • No patient data in billing, email, or the admin assistant
  • A transparent credit wallet — undo a top-up within 24 hours
A shield over a stack of layered records with padlocks and a node network
The billing and credits screen on a phone

That's the tour

Try it on a real workflow.

Start with a de-identified visit, pick a template, and review the generated draft before anything is signed.