# Universal-3 Pro Speech to Text

Transcribe audio and video to text with Universal-3 Pro Speech to Text from AssemblyAI through one Spiral API key. Transcribes audio or video into formatted text, steered by a written prompt or key terms, with speaker labels and PII redaction. It runs in the background and returns structured JSON. Pay as you go, with no subscription.

- Provider: AssemblyAI
- Category: Audio
- Runs: In the background
- Returns: json
- Released: 2026-02-03
- Tool id: `assemblyai` / `universal-3-pro`
- Page: https://spiral.ai/tools/assemblyai/universal-3-pro

## Pricing

$0.273 per hour of audio, per channel. Features, per hour: speaker labels $0.026, key terms $0.065, translation $0.078, speaker identification $0.026, custom formatting $0.039, action items $0.026, summaries $0.026 ($0.091 at medium effort), entity detection $0.104, sentiment $0.026, key phrases $0.013, topic detection $0.195, content moderation $0.195, profanity filter $0.013, PII redaction $0.104, audio redaction $0.065, medical mode $0.195, custom prompt $0.065.

## Limits

- Longest queue wait: 1 hour
- Run deadline: 30 minutes
- Longest run, end to end: 930 minutes
- Largest inline input: 1 MB
- Largest input file: 2 GB
- Largest output: 5 GB

## Input

| Field | Kind | Allowed | Default | Description |
| --- | --- | --- | --- | --- |
| `audio_url` (required) | URL or file |  | — | The URL of the audio or video file to transcribe. |
| `domain` | one of | medical-v1 | — | Enable domain-specific transcription models to improve accuracy for specialized terminology. |
| `prompt` | text |  | — | Provide natural language prompting of up to 1,500 words of contextual information to the model. |
| `topics` | list |  | — | This parameter does not currently have any functionality attached to it. |
| `punctuate` | true or false |  | true | Enable [Automatic Punctuation](https://www.assemblyai.com/docs/pre-recorded-audio), can be true or false |
| `redact_pii` | true or false |  | false | Redact PII from the transcribed text using the Redact PII model, can be true or false. |
| `format_text` | true or false |  | true | Enable Text Formatting. Universal-3 Pro always formats the text: `false` is refused. |
| `temperature` | number | 0 to 1 | — | Control the amount of randomness injected into the model's response. |
| `audio_end_at` | number (whole) |  | — | The point in time, in milliseconds, to stop transcribing in your media file. |
| `disfluencies` | true or false |  | false | Transcribe [Filler Words](https://www.assemblyai.com/docs/pre-recorded-audio/include-filler-words), like "umm", in your media file; can be true or false. |
| `multichannel` | true or false |  | false | Enable [Multichannel](https://www.assemblyai.com/docs/pre-recorded-audio/transcribe-multiple-audio-channels) transcription, can be true or false. |
| `custom_topics` | true or false |  | — | This parameter does not currently have any functionality attached to it. |
| `language_code` | JSON |  | — | The language of your audio file. |
| `content_safety` | true or false |  | false | Enable [Content Moderation](https://www.assemblyai.com/docs/content-moderation), can be true or false |
| `iab_categories` | true or false |  | false | Enable [Topic Detection](https://www.assemblyai.com/docs/speech-understanding/topic-detection), can be true or false |
| `language_codes` | list |  | — | The language codes of your audio file. Used for [Code switching](/speech-to-text/pre-recorded-audio/code-switching) One of the values specified must be `en`. |
| `redact_pii_sub` | one of | entity_name, hash | — | The replacement logic for detected PII, can be `entity_name` or `hash`. See [PII redaction](https://www.assemblyai.com/docs/pii-redaction) for more details. |
| `speaker_labels` | true or false |  | false | Enable [Speaker diarization](https://www.assemblyai.com/docs/pre-recorded-audio/label-speakers), can be true or false. Requires `punctuate` to be `true`. |
| `auto_highlights` | true or false |  | false | Enable [Key Phrases](https://www.assemblyai.com/docs/speech-understanding/key-phrases), either true or false |
| `custom_spelling` | JSON |  | — | Customize how words are spelled and formatted using to and from values. |
| `keyterms_prompt` | list |  | — | Improve accuracy with up to 200 (for Universal-2) or 1000 (for Universal-3.5 Pro) domain-specific words or phrases (maximum 6 words per phrase). |
| `speaker_options` | JSON |  | — | Specify options for [Speaker diarization](https://www.assemblyai.com/docs/pre-recorded-audio/label-speakers#set-a-range-of-possible-speakers). |
| `audio_start_from` | number (whole) |  | — | The point in time, in milliseconds, to begin transcribing in your media file. |
| `entity_detection` | true or false |  | false | Enable [Entity Detection](https://www.assemblyai.com/docs/speech-understanding/entity-detection), can be true or false |
| `filter_profanity` | true or false |  | false | Filter profanity from the transcribed text, can be true or false. |
| `redact_pii_audio` | true or false |  | false | Generate a copy of the original media file with spoken PII "beeped" out, can be true or false. |
| `speech_threshold` | number | 0 to 1 | — | Reject audio files that contain less than this fraction of speech. |
| `remove_audio_tags` | one of | all, speaker | — | Universal-3.5 Pro generates rich transcripts that can include inline annotations such as audio event markers and speaker cues. |
| `speakers_expected` | number (whole) |  | — | Tells the speaker label model how many speakers it should attempt to identify. |
| `language_detection` | true or false |  | — | [Automatic language detection](https://www.assemblyai.com/docs/pre-recorded-audio/language-detection) identifies the spoken language and routes the request to the best model. |
| `sentiment_analysis` | true or false |  | false | Enable [Sentiment Analysis](https://www.assemblyai.com/docs/speech-understanding/sentiment-analysis), can be true or false. Requires `punctuate` to be `true`. |
| `redact_pii_policies` | list |  | — | The list of PII Redaction policies to enable. See [PII redaction](https://www.assemblyai.com/docs/pii-redaction) for more details. |
| `speech_understanding` | JSON |  | — | Enable speech understanding tasks like [Translation](https://www.assemblyai.com/docs/speech-understanding/translation), [Speaker Identification](https://www.assemblyai.com/docs/speech-understanding/speaker-identification), and [Custom Formatting](https://www.assemblyai.com/docs/speech-understanding/custom-formatting). |
| `redact_static_entities` | JSON |  | — | A map of user-defined terms to redact, where each key is a redaction label and each value is a list of exact terms to match (e.g. `{ "INTERNAL_TOOL": ["Bearclaw", "Cubclaw"] }`). |
| `redact_pii_audio_options` | JSON |  | — | Specify options for [PII redacted audio](https://www.assemblyai.com/docs/pii-redaction#request-for-redacted-audio) files. |
| `redact_pii_audio_quality` | one of | mp3, wav | — | Controls the filetype of the audio created by redact_pii_audio. |
| `content_safety_confidence` | number (whole) | 25 to 100 | — | The confidence threshold for the [Content Moderation](https://www.assemblyai.com/docs/content-moderation) model. |
| `language_detection_options` | JSON |  | — | Specify options for [Automatic Language Detection](https://www.assemblyai.com/docs/pre-recorded-audio/language-detection). |
| `redact_pii_return_unredacted` | true or false |  | — | When set to `true`, returns the original unredacted transcript alongside the redacted one in the same response. |
| `language_confidence_threshold` | number | 0 to 1 | — | The confidence threshold for the automatically detected language. |

## Example

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "assemblyai",
    "model": "universal-3-pro",
    "input": {
      "audio_url": "<audio_url>"
    }
  }'
```

Try it in the Playground: https://spiral.ai/start?provider=assemblyai&model=universal-3-pro

## Common questions

### Universal-3 Pro Speech to Text

**How much does Universal-3 Pro Speech to Text cost?**

$0.273 per hour of audio, per channel. Features, per hour: speaker labels $0.026, key terms $0.065, translation $0.078, speaker identification $0.026, custom formatting $0.039, action items $0.026, summaries $0.026 ($0.091 at medium effort), entity detection $0.104, sentiment $0.026, key phrases $0.013, topic detection $0.195, content moderation $0.195, profanity filter $0.013, PII redaction $0.104, audio redaction $0.065, medical mode $0.195, custom prompt $0.065. You pay as you go from one Spiral balance, with no subscription, and a failed run is not charged.

**How do I use Universal-3 Pro Speech to Text through an API?**

Send `POST /v1/runs` with `"provider": "assemblyai"`, `"model": "universal-3-pro"` and the input fields on this page, using one Spiral API key. It runs in the background: poll the run, or name a webhook to be told when it finishes.

**Do I need an account with AssemblyAI to use Universal-3 Pro Speech to Text?**

No. Spiral runs it on its own AssemblyAI account and charges your Spiral balance.

**Can my AI agent use Universal-3 Pro Speech to Text?**

Yes. Agents find and run it through Spiral's MCP server, CLI or agent skill, with the same API key.

