Skip to content

Medical Speech Transcription

Every consultation, captured as structured text.

Convert doctor–patient conversations into diarized, timestamped transcripts with medical vocabulary handling — ready for search, documentation, and your EHR.

POST /v1/transcriptionsSynthetic example
response.json
{
  "id": "tr_demo_7d2jl5",
  "status": "completed",
  "language": "en",
  "duration_seconds": 28,
  "note": "Demo mode returns a fixed synthetic transcript regardless of audio_url.",
  "speakers": [
    {
      "id": "clinician",
      "label": "Doctor"
    },
    {
      "id": "patient",
      "label": "Patient"
    }
  ],
  "segments": [
    {
      "speaker": "clinician",
      "start": 0,
      "end": 4.2,
      "text": "Good afternoon. Can you describe when the symptoms started?"
    },
    {
      "speaker": "patient",
      "start": 4.6,
      "end": 9.8,
      "text": "The headache started yesterday evening, and I've felt nauseous since this morning."
    },
    {
      "speaker": "clinician",
      "start": 10.3,
      "end": 13.9,
      "text": "On a scale of one to ten, how strong is the pain right now?"
    },
    {
      "speaker": "patient",
      "start": 14.4,
      "end": 18.1,
      "text": "Around six. It gets worse when I look at bright light."
    },
    {
      "speaker": "clinician",
      "start": 18.7,
      "end": 22.9,
      "text": "Have you taken anything for it, or had a fever?"
    },
    {
      "speaker": "patient",
      "start": 23.3,
      "end": 27.6,
      "text": "I took ibuprofen last night. No fever that I noticed."
    }
  ]
}

Responsible use

Transcripts can contain errors. Clinicians should verify transcripts before relying on them for clinical documentation.

Capabilities

Clinical conversations, searchable and structured.

Status labels show what is in development today and what is planned.

  • Speech-to-text

    In development

    Consultation audio converted to clean text.

  • Speaker diarization

    In development

    Clinician and patient turns labelled automatically.

  • Timestamped transcripts

    In development

    Segment-level timing for review and navigation.

  • Medical vocabulary handling

    In development

    Drug names, anatomy, and clinical terms.

  • Multilingual transcription

    Planned

    English, Russian, Kazakh, Uzbek, and code-switching.

  • Real-time streaming

    Planned

    Live transcription during the consultation.

Workflow

How it fits your product

Key output fields
segments[]speakerstart / endlanguage
Typical host systems
Telemedicine platformsClinic workstationsEMR/EHR
  1. 1

    Audio submitted

    Pre-signed URL from your storage — files stay yours.

  2. 2

    Asynchronous processing

    Diarization, timestamps, and vocabulary handling.

  3. 3

    Webhook delivered

    transcription.completed with structured segments.

Demo

Explore a synthetic consultation

All examples on this site use synthetic data. No real patients, recordings, or inference requests are involved.

consultation_demo_001Demo
Synthetic transcript · no audio · simulated timeline
00:00 / 00:28
DoctorPatient

AKIRAMO

Structured clinical intake draft generated for clinician review.

chief_complaint
—
onset
—
associated
—
severity
—
aggravating
—
medication_taken
—
fever
—
Draft · pending reviewClinician verifies before use
Demo mode · syntheticProposed
POST/v1/transcriptions
Examples

Valid JSON

ResponseAwaiting request

No response yet.

Run the request to inspect the structured JSON response.

Responses are generated in your browser by deterministic demo rules over your request. They show the response schema — they are not results of a real inference request. Do not enter real patient data.

Your healthcare software deserves intelligence.

Integrate medical AI into the systems you already use.

AKIRAMO AI — Medical Intelligence InfrastructureDeveloped by ZhanCare.