Transcribe7 API

Speech-to-text infrastructure for your product

Real-time and recorded audio, speaker labels, 60+ languages and live translation — through one API. The same engine that runs Transcribe7.

€2 per audio hour · billed per second · no subscription

Live
Transcript
Speaker 1SL

Dober dan vsem, začnimo s pregledom projekta.

Speaker 2SL

Testiranje je končano, izdaja je pripravljena za petek.

Speaker 1SL

Odlično. Potem obvestimo stranke že ta teden.

AgentEN

Thanks for calling Northwind support, how can I help?

CustomerEN

My invoice shows two charges for the same order.

AgentEN

I've refunded the duplicate — you'll get an email shortly.

Speaker 1DE · EN

Ich schicke dir den pull request, it's ready for review.

Speaker 2SL · EN

Super, hvala — I'll check it after lunch.

Speaker 1DE

Perfekt, dann können wir morgen deployen.

Translation · EN

Good afternoon everyone, let's start with the project review.

Testing is finished, the release is ready for Friday.

Great. Then we notify customers this week.

Translation off for this stream — speaker labels and timestamps only.

Each word is tagged with its language, so switching languages mid-sentence is transcribed correctly.

Latency
while speaking
Speakers
labelled
Billing
per second
60+
languages, mixed freely
€2
per audio hour
1 GB
files, with webhooks
0
audio stored from live streams

Real-time

Text while people are still speaking

Stream audio over a WebSocket and receive words within moments. Each word arrives first as provisional and is confirmed once certain — show live text instantly, store only what's final.

  • Pause detection for voice assistants
  • Common formats or raw PCM
  • Short-lived keys for browsers and mobile apps
Live streamLive

Thanks for joining. Today we'll walk through the launch plan and the new pricing for Europe.

FinalProvisional

Speaker labels

Every word knows who said it

Speaker diarization labels each word with its speaker — live and in recordings. Build meeting minutes, interview transcripts and call analytics without guessing who spoke.

  • Exact timestamps per word
  • Real-time and files
  • Ideal input for AI summaries
board-meeting.m4a00:42
Speaker 1

Can we ship the update on Friday?

Speaker 2

Yes — QA signed off this morning.

Speaker 3

Then I'll tell the sales team today.

Languages & translation

60+ languages — and live translation

Hint the expected languages or let the API detect them per word. Add one-way translation for captions and audiences, or two-way for conversations between two languages.

  • Per-word language tags
  • One-way and two-way translation
  • Same price for every language
Spoken · Slovenian

Pogodba bo podpisana do konca meseca.

Lahko pošljete osnutek danes?

Translated · English

The contract will be signed by the end of the month.

Can you send the draft today?

Slovenščina English Deutsch Hrvatski Italiano Español Français Polski Українська 日本語 +50 more

Custom vocabulary

Your names and terms, spelled right

Pass product names, people and specialist terms with a short description of the topic. The engine uses it as context — no training, no fine-tuning.

  • Hundreds of terms per request
  • Topic description for context
  • Translation terms too
Without termsWe migrated cuber natives to the new trans cry be seven cluster in lube liana.
With your termsWe migrated Kubernetes to the new Transcribe7 cluster in Ljubljana.

Files & subtitles

Hours of recordings, done in minutes

Upload audio or video up to 1 GB, or point to a URL. A webhook tells your server when the transcript is ready — with timestamps, speakers, languages and translation.

  • Webhook or polling
  • Ready for subtitles, notes and search
  • Deleted automatically after 24 hours
interview.mp41 h 42 min
Completed · webhook delivered
1
00:00:01,200 --> 00:00:04,100
[Speaker 1] Good morning, everyone.

2
00:00:04,600 --> 00:00:08,900
[Speaker 2] Let's start with the quarterly results.

Use cases

Built for teams that work with conversations

Meeting & note-taking apps

Transcribe calls live, label speakers and hand clean text to your summary AI.

Contact centres

Every call searchable — for quality checks, coaching and compliance.

Media & podcasts

Subtitles and transcripts in minutes, in the original language and translated.

Events & accessibility

Live captions for conferences, lectures and town halls — in the audience's language.

Voice assistants

Fast text and pause detection for natural turn-taking.

Healthcare & legal

Specialist vocabulary spelled right; nothing from live streams is stored.

Privacy

Private by design, made in the EU

Conversations are sensitive. Real-time audio and text pass through and are never stored. Uploaded files and transcripts are deleted automatically 24 hours after they finish — or right away when you delete them.

  • Nothing kept from live streams
  • Files and transcripts deleted after 24 hours
  • EU company, GDPR-first

For developers

Switch in minutes — or let your AI do it

The API follows the common speech-to-text conventions: the same requests, responses and real-time protocol. Most apps switch by changing the address and the key. Our documentation is also published in a format AI coding assistants read directly.

Keys & projects

Separate keys per app or environment, revoke any time.

Team workspaces

Invite developers with owner, admin or developer roles.

Usage per project

Every second of audio, per day, key and project.

Private by design

Live audio never stored; files deleted after 24 h.

Pricing

One price. Every feature.

Prepaid credit, billed per second. No subscription, no minimum — credit never expires.

Pay as you go
€2 / audio hour
VAT included · real-time, files and translation at the same price
  • Real-time and file transcription
  • Speaker labels & language detection
  • One-way and two-way translation
  • Custom vocabulary & context
  • Webhooks, temporary keys, teams & projects
Get started
Audio per month 50 h
1 h1,000 h
Monthly cost €100

Large volumes or invoicing? support@transcribe7.com

Frequently asked questions

Is this the same engine as the Transcribe7 app?

Yes — the API runs on the same speech engine as the Transcribe7 apps and devices.

How is usage counted?

By seconds of audio processed. Real-time, files and translation all cost €2 per hour. Every charge is listed under Billing.

What happens when my credit runs out?

New streams and transcriptions are refused until you top up, and a live stream ends with a clear message when its credit is used up. Credit never expires.

Do you store my audio?

Real-time audio and text are never stored. Uploaded files and transcripts are deleted 24 hours after upload / completion, or earlier if you delete them.

Can my team work together?

Yes — workspaces with projects, roles (owner, admin, developer) and e-mail invitations.

Start building with Transcribe7

Sign in with your Transcribe7 account and make your first request in minutes.