sv
6 min läsning Integrations

API-integration

Fullständig VoxScriber API-dokumentation för utvecklare. Slutpunkter, autentisering och kodexempel.

The VoxScriber REST API lets you send audio by URL and receive the finished transcription in your system, without going through the site. It uses the same engines and charges the same cycles as transcriptions made in the dashboard.

Who can use it

The API is available starting on the Professional (R$ 39,90/mês) plan. Access is included for Professional, Premium, all Business plans (monthly and yearly), and equivalent legacy plans. If your account isn't eligible, key routes return error 403 PLAN_NOT_ELIGIBLE.

API keys

1

Create the key

Each account can have up to 20 active keys, each with an identifying name.
2

Copy the secret immediately

The key has the format vpt_live_... and is shown only once, at creation time. We store only a hash of it — it cannot be recovered later.
3

Store it securely

Keep it in an environment variable or a secrets vault. Never commit the key to versioned code.

Lost your key? There's no way to show it again. Revoke the old key and create a new one. Compromised keys should be revoked immediately.

There's no dedicated screen for creating personal keys in the dashboard yet. If your plan is eligible and you'd like to start using the API, open a ticket and the team will enable access with you.

Authentication

Every request carries the key in the Authorization header:

Authorization: Bearer vpt_live_SUA_CHAVE

A request without the header returns 401 MISSING_API_KEY.

Send an audio for transcription

The transcription endpoint is on the gateway api.voxscriber.com:

curl -X POST https://api.voxscriber.com/v1/transcriptions \
  -H "Authorization: Bearer vpt_live_SUA_CHAVE" \
  -H "Content-Type: application/json" \
  -d '{
    "audio_url": "https://exemplo.com/gravacoes/reuniao.mp3",
    "webhook_url": "https://seusistema.com/webhooks/vozparatexto",
    "formats": ["txt", "srt"]
  }'

Accepted fields in the body: audio_url, webhook_url, engine (ASSEMBLYAI, WHISPER or ELEVENLABS), language, formats (txt, json, srt, vtt, docx, pdf), metadata and idempotency_key. The result arrives at your webhook_url — see Webhooks and notifications.

Use idempotency_key with an identifier of yours (e.g., the recording ID in your system). If the same request is resent by mistake, the audio is not processed — or charged — twice.

Limits

LimitValue
Requests per key60 per minute
Audio downloaded by URLup to 500 MB (60 s timeout)
Audio via passthrough (Premium engine reads the URL directly)up to 5 GB
Active keys per account20

Security rules for URLs: https only, no redirects, and internal/private addresses are blocked. The audio URL must be publicly accessible.

The available export formats follow your plan's gate, same as the site — formats outside your plan are returned listed in exports_skipped.

Billing

The API debits the same cycles as a transcription made on the site, according to the chosen engine (Standard/Whisper: 1 cycle/min; Premium: 4 cycles/min; Ultra: 10 cycles/min). There is currently no monthly call quota or spending cap per key — control usage through your cycle balance. See Cycle system.

FAQ

I lost my API key. How do I recover it?

It can't be recovered — only the hash is stored in the database. Revoke the lost key and create a new one.

Does the API cost more than the site?

No. The cost in cycles is identical, based on the engine and the audio duration.

Can I upload the file directly instead of using a URL?

The main channel is via audio_url. If your audio doesn't have a public URL, contact support so we can evaluate the best flow for your case.

Is there OpenAPI documentation?

The public spec at /.well-known/openapi.json currently covers only public endpoints (statistics and shared transcriptions). For everything else, use this help center and support.

Still didn't resolve? Open a ticket — our team responds quickly.