Direct Audio/Video Transcriber + Diarized JSON avatar

Direct Audio/Video Transcriber + Diarized JSON

Pricing

from $6.00 / 1,000 transcription minutes

Go to Apify Store
Direct Audio/Video Transcriber + Diarized JSON

Direct Audio/Video Transcriber + Diarized JSON

Transcribe authorized direct audio/video files into diarized JSON with speaker segments and timestamps, plus SRT, VTT, and TXT. Platform pages are best-effort.

Pricing

from $6.00 / 1,000 transcription minutes

Rating

0.0

(0)

Developer

BRAINIALL Team

BRAINIALL Team

Maintained by Community

Actor stats

0

Bookmarked

1

Total users

1

Monthly active users

a day ago

Last modified

Categories

Share

Transcribe authorized direct audio or video file URLs into JSON, SRT, VTT, and TXT. The verified input path is a downloadable media file URL. Speaker diarization is optional.

Verified scope

  • Direct downloadable .mp3, .mp4, .wav, .ogg, .opus, .m4a, .webm, and .flac URLs
  • Files stored in an Apify key-value store when the URL is accessible to the run
  • JSON with timed words and segments
  • SRT, VTT, and TXT files saved to the run key-value store
  • Optional speaker labels such as SPEAKER_00 and SPEAKER_01
  • Explicit per-item errors so one failed source does not abort the remaining batch

Page URLs from video or social platforms are best-effort only. They may require cookies, authentication, or anti-bot challenges and are not guaranteed. For a reliable workflow, download media you are authorized to use and provide a direct file URL.

Quick start

{
"sources": [
"https://raw.githubusercontent.com/fasuizu-br/brainiall-asr-diarization-fixture/v1.1.0/fixture/transcreve-br-demo.mp3"
],
"language": "pt",
"diarization": true,
"output_formats": ["json", "srt", "vtt", "txt"],
"maxDurationMinutes": 2,
"integrationSource": "spring-ai-contract"
}

The example uses a version-pinned, first-party synthetic two-speaker fixture with an explicit permission file. Process only media you own or are authorized to use.

Structured transcription adapters

The output keeps aggregate text, detected language, duration, and timed speaker segments in the dataset row. This is useful when testing whether a framework adapter preserves diarized semantic data instead of collapsing the response to plain text.

An owner-controlled QA run on the pinned source produced eight timed segments and two anonymous speaker labels. That verifies this bounded runtime path only; it is not an external buyer, a WER/DER benchmark, or evidence of revenue.

Input

FieldTypeDefaultMeaning
sourcesarrayrequiredAuthorized direct media file URLs
languagestringautoSpoken language or automatic detection
diarizationbooleantrueAdd anonymous speaker labels
output_formatsarrayJSON, SRT, VTT, TXTRequested outputs
maxDurationMinutesinteger240Per-source duration cap
integrationSourcestringapify-storeNon-secret route tag for cohort attribution

integrationSource accepts 1–64 lowercase letters, numbers, ., _, or -. Never put a token, URL, email address, customer name, or personal data in this field. It identifies a route cohort, not a payer or buyer.

Output

Each source produces one dataset item. A successful item includes the source URL, attribution route, language, duration, charged minutes, transcript text, timed words or segments, and links to requested files. A failed item contains the source URL, attribution route, and a bounded error message.

Speaker labels are anonymous and may need editorial review. Subtitle cue boundaries and recognized names also need review before publication. The Actor does not certify accessibility, legal, or platform compliance.

Pricing and charging

The Actor uses Apify pay-per-event pricing by successful audio minute. Minutes are rounded up per source with a one-minute minimum. A failed source is recorded as an error item and is not charged as a successful transcription event.

Current creator rates depend on the user's Apify plan. Check the live Store pricing before estimating a batch; a published rate or plan can change.

Runtime boundaries

  • Default per-file cap: 240 minutes, configurable up to 1,440 minutes
  • Processing timeout: 30 minutes per source
  • A failed source does not abort other sources in the same run
  • Private, expired, or anti-bot-protected URLs can fail before transcription
  • Direct media URLs are the supported acquisition claim and automated-test path

Evidence boundary

A successful owner-controlled run proves the runtime path only. Runs, users, Store discovery, provisional developer revenue, invoices, and payout eligibility are separate stages. Revenue should be counted only after an independent buyer can be linked through a paid payout and reconciliation.

Support

Open an Actor issue with the run ID, input type, and a non-sensitive error summary. Do not include media content, credentials, customer data, or private URLs.