API

transcribe

Returns the full text plus time-coded segments. Stage the input with POST /v1/uploads first. ASYNCHRONOUS: the call returns 202 with a run id — poll GET /v1/runs/{id}, or register a webhook. Pass `language` when you know it: an explicit "ru" routes to a Russian-specialised engine that is an order of magnitude faster than the multilingual one, while "auto" has to use the multilingual model because the language is not known before decoding.

v3 beta server · 83 tokens

Price (server): 83 tokens per minute of audio, billed by the second, at least 83 per call; paid plans: −5%, rounded up. The exact figure for your file: POST /v1/tools/transcribe/estimate.

Call it

POST https://timbrica.com/v1/tools/transcribe/runs
Authorization: Bearer $TIMBRICA_KEY
Idempotency-Key: <your retry-safe key>
Content-Type: application/json

Input

FieldTypeRequiredNotes
audio object yes Staged input file — audio or video. Upload it with POST /v1/uploads (multipart `file`, or JSON {"url":"…"}). The audio track is used; video is ignored.
language string no ISO 639-1 code of the spoken language, or "auto" to detect it. Naming the language is both more accurate and, for "ru", much faster.
default: auto
model string no "auto" — our own engines (Russian-specialised for "ru", Whisper otherwise). "openai" — OpenAI gpt-transcribe: on our measurement about half the word errors of our Whisper on en, ko, ar, id, pt, de, es, fr and uk, priced per minute at its own rate (see the estimate). The audio is sent to OpenAI. Available on tembrica.com keys only (not timbrica.com), for those nine languages (or "auto").
one of: auto, openai
default: auto

Input file

Stage it with POST /v1/uploads (multipart `file`, or JSON {"url": "…"}), then pass {"file_id": "<id>"} in each listed input field.
file fields: audio

Errors it can return

validation_failed  too_long  insufficient_tokens  upstream_unavailable 

History

  • v3 — 2026-10-18 — The `model: "openai"` price is quoted only by the host that offers the model (tembrica.com); elsewhere the catalogue shows the base rate alone.
  • v2 — 2026-10-18 — Optional `model`: "openai" (OpenAI gpt-transcribe, intl only, own per-minute rate). The per-minute price now follows the site's.
  • v1 — 2026-08-09 — Initial release (server lane, per-minute pricing).

This contract cannot change without a new version and a line above: a test compares the published schema, pricing and limits against a committed lockfile and fails the build otherwise. Subscribe to /v1/changelog to be told when something moves.