# Speechmatics

Speech-to-text API that transcribes recorded files and live audio streams into text with timings and speaker labels.

- Provider: Speechmatics (https://www.speechmatics.com)
- Category: Transcribe audio
- Type: Hosted API
- Pricing model: Usage-based credits, billed per second at an hourly rate [2]
- Free tier: $100 credit grant, no card [3]
- Authentication: API key as bearer token; temporary JWTs for client-side use [5]
- Connect via: REST API (batch), WebSocket (realtime), Python, JavaScript, .NET and Rust SDKs [13]
- Read from the provider's pages on 2026-11-04
- Page: https://supermarket.dev/tools/transcribe-audio/speechmatics

## Pricing

### Billing
- Pricing model: Pay as you go, credit packs or subscription; credits at $1 each, billed per second [2]
- Free tier: $100 credit grant for new sign-ups [3]
- Card required for free tier: No [3]
- Minimum spend: Not published
- Billing period: Monthly; usage beyond granted credits is charged on the first of the next month [3]
- Credit expiry: Sign-up bonus, credit-pack and annual subscription credits last 12 months; monthly subscription credits last 2 months [2]

### Prices
Rates as listed per hour of audio; billed to the exact second. Linden 1 realtime shows two prices on the page, $0.20 and $0.15, and the page does not say which applies.
| Model | Pre-recorded ($/hour) | Real-time ($/hour) |
|---|---|---|
| Melia 1 (multilingual) | $0.12 | $0.20 |
| Enhanced | $0.38 | $0.40 |
| Standard | $0.23 | $0.23 |
| Oak 1 (medical) | $0.15 | $0.23 |
| Linden 1 (voice agents) | Not offered | $0.20 and $0.15 both shown |
Source [1]

### Plans
| Plan | Details |
|---|---|
| Free | $100 credit grant, no card; 2 concurrent realtime sessions; API access pauses when credits run out |
| Pro | Pay as you go on a card; 50 concurrent realtime sessions; volume and model-training discounts |
| Enterprise | Contract billing; custom limits; on-prem and on-device deployment |
Source [3]

### Other charges
- Model-training discount: 33% off speech-to-text rates if you allow audio and transcripts to be used to improve models [1]
- Volume discount: Up to 25% off, applied automatically above 500 hours a month per product type [3]
- Add-ons, pre-recorded: Translation $0.49/hr; Chapters $0.30/hr; Topics $0.15/hr; Summaries $0.09/hr; Sentiment $0.09/hr [1]
- Text to speech: $0.011 per 1,000 characters [1]

## Tools

### MCP server
- Official MCP server: No [19]
- Address or package: Not published

### API
| Endpoint | What it does |
|---|---|
| POST /v2/jobs | Create a batch transcription job |
| GET /v2/jobs | List jobs |
| GET /v2/jobs/{id} | Get job details and progress |
| GET /v2/jobs/{id}/transcript | Get the transcript |
| DELETE /v2/jobs/{id} | Delete a job and its data |
| GET /v2/jobs/{id}/usage | Get usage statistics |
| wss://eu.rt.speechmatics.com/v2 | Realtime transcription over WebSocket |
| wss://global.rt.speechmatics.com/v2/agent | Agent STT, turn-based transcription for voice agents |
Source [15]

### Features
- Default model: Not published
- Languages: 55+ languages; Enhanced and Standard need a language chosen, Melia 1 and Oak 1 switch automatically [11]
- Speaker labels: Speaker and channel diarization (speaker diarization on batch only for Melia 1 and Oak 1) [11]
- Word timestamps: Yes, word-level timings in batch and realtime [11]
- Live streaming: Yes, over WebSocket [15]
- Max file size: 1 GB in the request body; larger files by URL [9]
- Audio input: Batch: wav, mp3, aac, ogg, mpeg, amr, m4a, mp4, flac; realtime: raw PCM or file [12]

## Authentication

### Methods
| Method | Used for | How |
|---|---|---|
| API key | Batch, realtime and Agent STT, server-side | `Authorization: Bearer $API_KEY` header |
| Temporary key (JWT) | Client-side realtime and Agent STT calls | `?jwt=$TEMP_KEY` query parameter on the WebSocket URL |
| Management token | Workspace admin, such as creating projects and API keys | Management API |
Source [5]

### Getting access
- Who creates the account: A person signs up in the portal; agent self-sign-up is not stated [5]
- Where to get a key: Portal, Settings > API Keys; Management API with a management token [5]
- Other requirements: Enterprise customers may need to contact support for keys; no card for the free plan [14]

## Setup

### Addresses
- REST API: `https://eu1.asr.api.speechmatics.com/v2` (also `us1` and `au1`) [5]
- MCP server: Not published
- OpenAPI spec: Not published
- Docs for agents: `https://docs.speechmatics.com/llms.txt`; every docs page also has a `.md` copy [15]
- Status page: `https://status.speechmatics.com` (HTTP 200) [19]

### Install
| What | Command |
|---|---|
| Python batch SDK | `pip install speechmatics-batch` |
| JavaScript batch SDK | `npm install @speechmatics/batch-client` |
Source [14]

## Limits

### Rate limits
- Batch job creation: 10 new jobs per second recommended; HTTP 429 beyond that [9]
- Batch status requests: 50 per second recommended [9]
- Concurrent realtime sessions: Free 2, Pro 50, Enterprise custom [10]

### Other limits
- Concurrent batch jobs: 20,000 [9]
- Realtime session length: Ends at 48 hours, after 1 hour with no audio, or after 3 minutes with no audio or ping/pongs [10]
- Custom dictionary size: 20,000 items; more is rejected [9]

## Privacy

### Your data
- Data location: EU1, US1 or AU1, chosen by the endpoint called [6]
- Retention: Batch: stored 7 days then deleted; realtime: not stored [7]
- Used for training: Only if you opt in to the model-training discount; Enterprise says no logging by default [3]
- Zero data retention: Realtime does not store audio or transcripts [10]

### Compliance
- Certifications: ISO/IEC 27001:2022, SOC 2 Type II, GDPR, HIPAA [7]
- HIPAA: Stated as compliant [8]

## Sources

1. Pricing: https://www.speechmatics.com/pricing
2. Pricing FAQ: https://www.speechmatics.com/pricing
3. Plans: https://docs.speechmatics.com/administration/plans.md
4. Billing: https://docs.speechmatics.com/administration/billing.md
5. Authentication: https://docs.speechmatics.com/get-started/authentication.md
6. Regions: https://docs.speechmatics.com/administration/regions.md
7. Security and compliance: https://docs.speechmatics.com/administration/security-and-compliance.md
8. Security: https://www.speechmatics.com/security
9. Limits, batch: https://docs.speechmatics.com/speech-to-text/batch/limits.md
10. Limits, realtime: https://docs.speechmatics.com/speech-to-text/realtime/limits.md
11. Models: https://docs.speechmatics.com/speech-to-text/models.md
12. Batch input: https://docs.speechmatics.com/speech-to-text/batch/input.md
13. SDKs: https://docs.speechmatics.com/integrations-and-sdks/sdks.md
14. Batch quickstart: https://docs.speechmatics.com/speech-to-text/batch/quickstart.md
15. llms.txt: https://docs.speechmatics.com/llms.txt
16. Create a new job: https://docs.speechmatics.com/api-ref/batch/create-a-new-job.md
17. GitHub org: https://github.com/speechmatics
18. Status: https://status.speechmatics.com
19. Checked by supermarket.dev