Speech to Text
Category: AI & ML
Node id: speech-to-text
Standards: HIPAA
Transcribes audio files to text using ElevenLabs Scribe v1 (speaker diarization, 99 languages, timestamped segments) or Gemini on Vertex AI (provider: gemini)
What it does
- Audio transcription via ElevenLabs Scribe v1 (state-of-the-art accuracy)
- 99 language support with automatic language detection
- Speaker diarization with configurable speaker count
- Timestamped segments at word or sentence level
- Supports audio URLs or inline audio data
- Handles files up to 3GB / 10 hours duration
Use cases
- Transcribe physician-patient encounter recordings for SOAP note generation
- Convert telemedicine audio to searchable clinical documentation
- Process Docline consultation recordings for clinical workflow pipelines
- Multi-speaker clinical meeting transcription with speaker identification
Runtime
- Performance tier:
slow
Input ports
| Name | Type | Required | Description |
|---|---|---|---|
audio_data | text | No | Base64-encoded or raw audio data |
audio_url | text | No | URL to fetch audio file from |
Output ports
| Name | Type | Required | Description |
|---|---|---|---|
transcript | text | No | Full transcription text |
segments | json | No | Timestamped segments with speaker labels |
speakers | json | No | Speaker information from diarization |
language | text | No | Detected or specified language code |
duration | number | No | Audio duration in seconds |
success | boolean | No | Whether transcription succeeded |
Settings
Basic
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Provider | provider | enum | No | elevenlabs | elevenlabs, gemini | Transcription provider: ElevenLabs Scribe, or Gemini on Vertex AI (Application Default Credentials, no API key) |
| Gemini Model | model | string | No | Gemini model for provider gemini (falls back to GEMINI_MODEL) | ||
| Model | model_id | enum | No | scribe_v1 | scribe_v1, scribe_v1_experimental | Transcription model |
| Language | language | string | No | Language code (auto-detect if empty) | ||
| Detect Speakers | detect_speakers | boolean | No | true | Enable speaker diarization | |
| Number of Speakers | num_speakers | number | No | 0 | min 0; max 20 | Expected speaker count (0 = auto) |
Auth
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| ElevenLabs API Key | api_key | secret | No | API key for ElevenLabs (provider elevenlabs only) | ||
| GCP Project | gcp_project | string | No | Vertex AI project for provider gemini (falls back to GEMINI_PROJECT / GOOGLE_CLOUD_PROJECT) | ||
| GCP Location | gcp_location | string | No | Vertex AI location where the recording is processed, provider gemini (falls back to GEMINI_LOCATION; no implicit global default) |
Advanced
| Setting | Name | Type | Required | Default | Allowed values | Description |
|---|---|---|---|---|---|---|
| Audio MIME Type | mime_type | string | No | Audio type sent to Gemini (e.g. audio/wav, audio/mpeg); detected from the input when empty | ||
| Vertex Base URL | base_url | string | No | Vertex AI endpoint override for provider gemini (private endpoints, testing) | ||
| API URL | api_url | string | No | https://api.elevenlabs.io/v1/speech-to-text | ElevenLabs STT endpoint | |
| Timestamp Granularity | timestamp_granularity | enum | No | segment | segment, word | Timestamp level |
| Timeout (seconds) | timeout | number | No | 300 | min 10; max 600 | Request timeout for large files |
Errors
| Code | Description |
|---|---|
API_KEY_REQUIRED | ElevenLabs API key not configured (provider elevenlabs) |
VERTEX_NOT_CONFIGURED | No Vertex AI project, location or model for provider gemini (gcp_* / model config or GEMINI_* env) |
NO_AUDIO_INPUT | Neither audio_data nor audio_url was provided |
AUDIO_FETCH_FAILED | Failed to download audio from the provided URL |
STT_REQUEST_FAILED | ElevenLabs API call failed |
EMPTY_TRANSCRIPT | Transcription succeeded but returned no text |
Runtime notes
- Provide either audio_data (base64/raw bytes) or audio_url (fetched automatically) — not both
- transcript is the concatenated full text; segments contains per-utterance timestamps and speaker IDs
- Speaker diarization is enabled by default; set detect_speakers=false to skip
- Timeout defaults to 300s (5 min) to accommodate large audio files up to 3GB / 10 hours
- success=false with an empty transcript when the audio contains no recognizable speech
See Build any workflow for how to put nodes together and Ports, types and data flow for how port types connect.
Something unclear? Send documentation feedback by email.