Transcribe with the best speech engines. Deliver results that are ready to use.
Integrate once and combine Whisper, Deepgram, AssemblyAI, Soniox, ElevenLabs, Gladia, Voxtral, iFlytek and Alibaba Qwen in a single infrastructure layer. Start with competitive raw transcription, scale with confidence and, when the job calls for more, turn audio into documents, templates, summaries and automations ready to move forward.
POST /api/v1/transcribe
curl -X POST \
https://api.transcriwise.com/api/v1/transcribe \
-H "Authorization: Bearer tw_..." \
-F "file=@audio.mp3" \
-F "mode=APOSTILA" \
-F "webhook_url=https://app.exemplo.com/hook"Example based on the implemented public endpoint. Replace the key, file, and webhook.
Create the key
Generate a key with the required scope in the developer panel.
Upload the file
Use multipart/form-data on the transcription endpoint.
Receive the webhook
Provide a public URL for the asynchronous callback.
Read the result
Use the job_id for status and artifacts allowed by the plan.
Why use it
Integrate once. Choose the path by context, not trial and error.
One unified API normalizes authentication and responses across enabled providers without requiring a different key for each call.
No vendor lock-in
The premium lives in the final delivery
Scale, governance and compliance
API features
Everything you need in a single endpoint.
From raw transcription to professional documents, with AI add-ons and export in multiple formats.
Batch and live transcription
Record and upload, or transcribe in real time over WebSocket. Multiple providers support live with autorouter and pinning.
Routing and availability
The autorouter considers language, cost, capabilities, and providers enabled at call time.
Speaker diarization
Know who said what when the selected provider and lane support diarization.
Code-switching (language switching)
Recognizes when the speaker changes language mid-sentence. Soniox, Gladia and AssemblyAI.
Simultaneous translation
Soniox: two-way translation in 60+ languages. iFlytek: Chinese→English with TTS.
Professional documents (v2)
Generate minutes, legal drafts, expert reports, meeting notes and more — billed per document.
Summary, mind map and chat
Executive summary, mind map, flashcards, quiz and chat over the transcript.
DOCX, PDF and SRT export
Export with a visual template, generate SRT/VTT captions, and apply custom templates.
Webhooks and callback
Get a POST when the job completes or fails. No polling.
BYO Keys
Use your own provider key. Costs go straight to your provider account on the Enterprise track.
Normalized payload + raw
Standardized response by default. Raw provider JSON with include_raw=true.
Cost caps and rate limits
Automatic protection against accidental usage. Configurable per plan.
V2 Documents
From audio to a finished document in 4 steps.
Pick the document type, apply your template, refine with AI and generate study material — all via API.
Document type
Custom template
AI refinement
Study material
Plans
The API ships with Pro. Or use it without the app.
Transcriwise users get API access from Pro up. If you only need the API, you can start free with the Sandbox and scale with PAYG.
API included in the current app
The API is available from Pro up. Transcription uses your plan quota. Professional documents are available on the higher plans.
| App plan | API access | Rate limit | Cost cap | Coverage |
|---|---|---|---|---|
| Starter | Not included | - | - | App only, no API access |
| Pro | Included | 60 req/min | US$ 15/day · US$ 150/month | API included + plan-compatible catalog + webhooks |
| Business | Included | 100 req/min | US$ 50/day · US$ 500/month | Everything in Pro + more volume + priority support |
| Enterprise | Included | 200 req/min | Custom | V1 + V2 + BYO Keys + SLA and dedicated lanes |
Sandbox
Free
Free sandbox to test V1 raw without subscribing to the app.
- 5 req/min
- US$ 5/day · US$ 10/month
- Recordings only
- Automatic routing (provider=auto)
Build
PAYG
To integrate V1 raw with pay-as-you-go pricing and the full catalog.
- 30 req/min
- US$ 25/day · US$ 100/month
- Recordings and live
- Autorouter + enabled catalog + webhooks
Scale
Best balanceUS$ 99/m
For production volume, V2 documents and better support.
- 120 req/min
- US$ 100/day · US$ 1000/month
- V1 raw + V2 documents
- Same full provider catalog + priority support
Enterprise
Contact sales
For compliance, SLA, BYO Keys, dedicated lanes and regional routing.
- 300 req/min
- Custom
- SLA + BYO Keys
- Dedicated lanes, regional routing and self-hosted
Hourly pricing (transcription)
Automatic routing considers language, cost, capabilities, and availability. When the plan allows it, you can pin a provider and use its specific table; app plans use their respective quotas.
To generate professional documents (minutes, legal drafts, expert reports), use the documents API with per-document pricing.
| Autorouter lane | USD | BRL | Coverage |
|---|---|---|---|
| Batch Standard Auto | US$ 0.29/h | R$ 1,65/h | Soniox, Whisper Turbo, Gladia, Voxtral, AssemblyAI U-2, Qwen3-ASR |
| Batch Premium Auto | US$ 0.79/h | R$ 4,50/h | AssemblyAI U-3 Pro, Deepgram, ElevenLabs, Whisper Large V3 |
| Live Standard Auto | US$ 0.69/h | R$ 3,93/h | Soniox, AssemblyAI U-2, ElevenLabs, Gladia |
| Live Premium Auto | US$ 1.29/h | R$ 7,35/h | AssemblyAI U-3 Pro, Deepgram, Soniox |
Per-provider pricing — recordings
Pricing when you pin the exact provider instead of using the autorouter. Explicit selection requires provider pinning in the plan.
| Provider / model | USD | BRL | Availability |
|---|---|---|---|
| Whisper V3 Turbo | US$ 0.29/h | R$ 1,65/h | Pro+/Build+ |
| Soniox Async v5 | US$ 0.18/h | R$ 1,03/h | Pro+/Build+ |
| Gladia Batch | US$ 1.09/h | R$ 6,21/h | Pro+/Build+ |
| Voxtral Batch | US$ 0.24/h | R$ 1,37/h | Pro+/Build+ |
| AssemblyAI U-2 | US$ 0.24/h | R$ 1,37/h | Pro+/Build+ |
| AssemblyAI U-3 Pro | US$ 0.36/h | R$ 2,05/h | Pro+/Build+ |
| Whisper Large V3 | US$ 0.39/h | R$ 2,22/h | Pro+/Build+ |
| Deepgram Nova-3 | US$ 0.89/h | R$ 5,07/h | Pro+/Build+ |
| ElevenLabs Scribe v2 | US$ 0.69/h | R$ 3,93/h | Pro+/Build+ |
| Qwen3-ASR | US$ 0.19/h | R$ 1,08/h | Pro+/Build+ |
| SenseVoice self-hosted | US$ 0.16/h | R$ 0,91/h | Pro+/Build+ |
| Paraformer-v2 | US$ 0.16/h | R$ 0,91/h | Pro+/Build+ |
| Alibaba SenseVoice hosted | US$ 0.24/h | R$ 1,37/h | Pro+/Build+ |
| iFlytek LFASR Basic | US$ 1.19/h | R$ 6,78/h | Pro+/Build+ |
| iFlytek LFASR LLM | US$ 1.60/h | R$ 9,12/h | Pro+/Build+ |
Per-provider pricing — live
Pricing for real-time transcription. Live routing and explicit selection depend on plan features.
| Provider | USD | BRL | Availability |
|---|---|---|---|
| Soniox live | US$ 0.29/h | R$ 1,65/h | Pro+/Build+ |
| AssemblyAI U-2 live | US$ 0.29/h | R$ 1,65/h | Pro+/Build+ |
| Gladia live | US$ 1.29/h | R$ 7,35/h | Pro+/Build+ |
| AssemblyAI U-3 Pro live | US$ 0.79/h | R$ 4,50/h | Pro+/Build+ |
| Deepgram live | US$ 0.99/h | R$ 5,64/h | Pro+/Build+ |
| Qwen3-ASR realtime | US$ 0.49/h | R$ 2,79/h | Pro+/Build+ |
| iFlytek realtime | Custom | Custom | Scale+/Enterprise |
| Whisper / SenseVoice live dedicated | Custom | Custom | Enterprise |
Professional documents (v2)
Flat price per fixed-format document, separate from transcription cost. Long-form content and transcript-only use duration brackets (tables below). Available from the Scale plan (API-only) or Enterprise (app).
| Category | USD | BRL | Scope |
|---|---|---|---|
| General | US$ 0.50/doc | R$ 2,85/doc | meeting_notes, podcast, interview, call (flat price) |
| Healthcare | US$ 1/doc | R$ 5,70/doc | consulta_medica and clinical flows |
| Legal | US$ 1.50/doc | R$ 8,55/doc | legal_draft, deposition, court_hearing, court_ruling, notarial, pericia |
Formatted document by duration
Long-form content (lecture, study, webinar, simulation, fidelity): flat price per duration bracket, with transcription + AI formatting included.
| Audio duration | USD | BRL |
|---|---|---|
| Up to 30 min | US$ 1.99/doc | R$ 11,35/doc |
| Up to 60 min | US$ 3.99/doc | R$ 22,75/doc |
| Up to 120 min | US$ 7.99/doc | R$ 45,55/doc |
| Up to 240 min | US$ 15.99/doc | R$ 91,15/doc |
Above 240 min: US$ 0.07/min · R$ 0,40/min per extra minute.
Transcript-only by duration (~half price)
raw_transcript: transcribed text without AI formatting. Costs about half of the formatted document, keeping the same duration-bracket model.
| Audio duration | USD | BRL |
|---|---|---|
| Up to 30 min | US$ 0.99/doc | R$ 5,65/doc |
| Up to 60 min | US$ 1.99/doc | R$ 11,35/doc |
| Up to 120 min | US$ 3.99/doc | R$ 22,75/doc |
| Up to 240 min | US$ 7.99/doc | R$ 45,55/doc |
Above 240 min: US$ 0.04/min · R$ 0,23/min per extra minute.
Duration brackets apply to long-form content types (lecture, study, webinar, simulation, fidelity) and to raw_transcript (transcript only). They are also available on PAYG/one-off, with no plan.
Coverage
Autorouter and provider pinning are different capabilities.
The autorouter uses providers enabled for Batch or Live. Explicit provider and model selection requires provider pinning; throughput, documents, support, and infrastructure also vary by plan.
Catalog in autorouter; pinning according to plan
- Whisper V3 Turbo and Large V3
- AssemblyAI (U-2 and U-3 Pro)
- Deepgram Nova-3
- ElevenLabs Scribe v2
- Soniox v4 (batch and live)
- Gladia (batch and live)
- Voxtral (Mistral)
- iFlytek (batch and live, zh→en translation)
- Alibaba Qwen3-ASR, Paraformer-v2, SenseVoice
- SenseVoice self-hosted
What changes per plan
- Sandbox: autorouter only (provider=auto)
- Build+: autorouter + provider pinning + webhooks
- Scale+: everything in Build + v2 documents + priority support
- Enterprise: BYO Keys, SLA, dedicated regions and custom lanes
Add-ons
Extra capabilities priced by the value they deliver.
Diarization, translation, summarization and more — priced proportionally to the provider and model used.
Diarization
+25%Who said what, with speaker identification.
PII redaction
+25%Masking of sensitive data (LGPD/GDPR).
Keyterms / prompting
+15%Vocabulary control and better terminology adherence.
Translation
+30%On top of the chosen provider cost: Gemini, GPT, DeepL, Azure, Google NMT or Google TLLM.
Summary and AI
+25%On top of the LLM cost used for summary, mind map, chat and entities.
Speech Understanding
Advanced analyses on top of the transcript, varying by provider.
AssemblyAI
Each feature: US$ 0,006/min (US$ 0,36/h) AAI cost.
Gladia
Everything included in the Gladia base price — no extra cost.
ElevenLabs
Official ElevenLabs surcharge.
Two experiences, one infrastructure.
The right provider for every situation.
Recordings: Soniox and Gladia for multilingual flows. Whisper and Voxtral for batch processing. Deepgram for specialized profiles. iFlytek and Alibaba for Chinese.
Live: Soniox for captions and simultaneous translation. Gladia for multilingual meetings with code-switching. Deepgram for live-audio profiles. iFlytek for Chinese→English in real time.
Frequently asked questions
Everything you need to know before integrating.
Straight answers about providers, pricing, formats and scalability.
Can I choose the provider?
Does the response include the provider raw data?
Can I generate formatted documents via the API?
What is the best provider for Brazilian Portuguese?
Can I use my own provider key?
Get started
Start free with the Sandbox. Scale when you need to.
Start in the Sandbox if you only want V1 raw. Use Build for early production. If you are already in the app, Pro is the most efficient way to integrate without buying another stack. Move up to Business, Scale or Enterprise when the bottleneck becomes throughput, support, documents or governance.