Skip to main content
Ask Brigid
Request early accessExplore the sample practiceTry it

Launch-scope dictation

EU speech-to-text,with the clinician in control.

Brigid's launch boundary is deliberately narrow: speech becomes raw text on the configured Google EU route, and a note or letter can then be drafted from it for you to review, correct and sign. AI that diagnoses or triages is off.

Request early access
Built in Ireland for private practiceEvery clinical word is authored by a clinicianEarly access for founding practices

What is enabled

Dictation support—not clinical generation.

Hold to dictate

Start dictation deliberately from the field you are editing. The app does not treat ambient conversation as permission to record.

Raw transcript

Speech-to-text returns words, not an assessment, plan, diagnosis, prescription or medical recommendation.

EU route

Release-candidate source selects Google Speech-to-Text Chirp 3 at the configured EU location and fails closed rather than using a US fallback.

Accuracy requires review

Names, medicines, doses, laterality and negation must be checked by the clinician before text is accepted.

Clinician-authored record

The treating professional decides what belongs in the note and signs it. Templates may structure the page, but they do not supply clinical judgement.

Evidence before enablement

Clinic instructions, Google terms/configuration, notice, retention, deletion and production logs must be filed before consultation recording is enabled.

Clinic rollout

Prove the boundary before patient use.

  1. Step 1

    Approve the purpose

    The clinic records why dictation or recording is necessary and selects the correct lawful basis, notice and retention rule.

  2. Step 2

    Test the EU route

    Staging logs prove the provider, model and EU location, including a negative test showing there is no non-EU fallback.

  3. Step 3

    Train and rehearse

    Clinicians practise review, correction, cancellation, deletion and downtime handling before the feature touches patient data.

Questions clinicians ask

What the launch route does—and does not do.

Does it understand Irish accents?
Dictation uses Google Cloud Speech-to-Text Chirp 3 on the EU route. Accuracy varies by speaker, microphone, terminology and environment, so each clinic must test it and clinicians must correct the text before signing.
What happens to the audio?
The approved source route sends audio to Google Cloud Speech-to-Text at the configured EU location and has no ElevenLabs or US medical-model fallback. Production region, logging, contract and deletion evidence is still outstanding.
Can the transcript file itself into the record?
No. Raw text is returned for clinician review. The clinician decides what belongs in the record and remains responsible for accuracy, context and sign-off.
Can recording be enabled automatically?
No. Recording follows the clinic’s approved purpose, lawful basis and notice or permission procedure, and the original audio is deleted within 24 hours after the clinician accepts the note, and in any case within seven days.

Launch boundary

Raw text. Clinician judgement.

EU
Configured Google Speech-to-Text route
None
Clinical decisions made by AI
Fail closed
No ElevenLabs or US medical-model fallback
Human
Review and sign-off remain mandatory

Test dictation without changing who writes the record.

Request a controlled demo using synthetic data. Patient use remains gated by the clinic's approved procedure and production evidence.

Request early access

Blog

From the notes

All notes

Explore more

Clinical records

Staff-authored records, templates and launch-scope dictation.

Compliance status

Current boundaries, provider evidence and the evidence still outstanding.

Sub-processors

The current provider inventory and evidence status.

Pricing

Published early-access plans and included administrative tools.