Arabic transcription API

Kalemio is an Arabic speech-to-text API for products that need more than a block of text. Send an Arabic audio or video file, poll an asynchronous job, and download a transcript with word timing, speaker labels, JSON, or SRT.

Arabic as it is spoken

Use the API for Modern Standard Arabic and spoken dialects. The result aims to keep the wording in the recording instead of automatically rewriting dialect into formal Arabic. Review names, music, overlapping voices, and important text before publishing.

What the API returns

  • Full transcript text and an array of words.
  • Start and end timestamps for every word.
  • Speaker labels from automatic detection or a known speaker count.
  • JSON for search, playback, and custom interfaces, plus a ready-to-import SRT file.

A first request

curl "$KALEMIO_API_BASE/v1/transcriptions" \
  -H "Authorization: Bearer $KALEMIO_API_KEY" \
  -H "Idempotency-Key: recording-2026-001" \
  -F '[email protected]' \
  -F 'language=ar' \
  -F 'speakers=auto'

The upload returns 202 Accepted and a job ID. Poll that job, then request its JSON result or SRT download. Keep API keys on your server and grant each key only the permissions it needs.

Limits, billing, and retention

Send MP3, M4A, PCM WAV, MP4, or MOV files up to 20 MiB and 10 minutes each. Two jobs can be active per account. Prepaid plans measure uploaded media duration rounded up to the next second and do not use automatic overages. An optional, separate usage-based subscription is billed in USD at US$0.09 per processed audio minute, rounded up to one minute, after a successful job. Usage-based accounts do not need prepaid minutes. Working audio is deleted after processing and results are available for 24 hours, so download what you need.

Build captions or searchable transcripts

Use SRT when you want a subtitle file for an editor or player. Use the word timestamps in JSON to highlight speech during playback, jump from search results to the right moment, or build your own caption layout. Read the API documentation for response fields, errors, and polling examples.

Privacy and access

Submitted recordings are not used to train models. Authentication uses bearer API keys, and the API console lets you create, inspect, and revoke keys. See the API privacy policy for retention and service-provider details.

Get an API key Read the documentation

Need a subtitle file rather than an integration? Arabic Subtitles is a separate no-code product, with guides for dialect subtitles, YouTube, and Arabic SRT files.