Jobs API
The Jobs API is an asynchronous (batch) speech-to-text API for prerecorded audio and video files. You create a job from a file URL or an upload, and the API returns a job ID right away. You then receive the transcript by webhook, follow it as a stream, or poll the job. Use it for long recordings and for output beyond plain text, such as speaker diarization, word-level timestamps, and SRT or WebVTT subtitles.
For a short upload whose transcript you want in the same response, use the Edge API. To pay for each job from a crypto wallet instead of with an API key, use the x402 API. How to Transcribe compares them.
Base URLs
| Host | Used for |
|---|---|
https://api.speechischeap.com | Creating, getting, and canceling jobs, and health checks |
https://upload.speechischeap.com | Uploading a file. The host forwards the file to the API and returns the same job |
Authentication
Requests that create, upload, or cancel a job need an API key, sent as a Bearer token:
Authorization: Bearer YOUR_API_KEY
API keys start with sk_. Sign up on the pricing page, and your first key arrives by email. You create, rename, and revoke keys in the Dashboard. How to Transcribe walks through the sign-up. To test a key before you use it, check it.
Getting a job and the health checks need no key. Anyone with a job ID can read that job, so keep job IDs private.
Endpoints
| Method | Endpoint | Page |
|---|---|---|
POST | https://api.speechischeap.com/v2/jobs/ | Create a Job from a URL |
POST | https://upload.speechischeap.com/v2/jobs/ | Upload a File |
GET | https://api.speechischeap.com/v2/jobs/:job_id | Get a Job |
DELETE | https://api.speechischeap.com/v2/jobs/:job_id | Cancel a Job |
GET | https://api.speechischeap.com/v2/jobs/health | API Health |
GET | https://api.speechischeap.com/v2/jobs/health/inference | API Health |
GET | https://upload.speechischeap.com/health | API Health |
GET | https://api.speechischeap.com/v2/jobs/auth | API Health |
Requests that create a job also work without the trailing slash. Don't add a slash to the other paths. A request to /v2/jobs/:job_id/ returns 404 Not Found.
Webhooks, streaming, and subtitles have no endpoint of their own. You turn them on with parameters when you create the job. See Webhooks, Stream Progress, and Output and Add-ons.
Job States
A job starts as PENDING and ends in one of the other three states.
| Status | Meaning |
|---|---|
PENDING | The job is queued or being transcribed. |
COMPLETED | The transcript is ready. |
FAILED | The job couldn't be transcribed. You aren't charged. |
CANCELED | The job was canceled. |
How Results Arrive
You can use any of these together:
- A webhook sends the result to your
webhook_urlonce, when the job ends. - A stream sends progress and segments while the job runs. Set
can_stream_outputwhen you create the job. - Get a Job returns the job by its ID. Request it once, or in a loop until the job ends.
The result is JSON, or SRT or VTT subtitles. See Output and Add-ons.
Billing
You pay only when a job completes. A job that fails or is canceled before it completes costs nothing. The API rounds each file up to a whole minute, separately for the transcription and for each add-on you turn on. See pricing for the rates. Jobs paid with x402 are paid from your wallet instead. See Payments and Refunds.
Related Pages
- Input Requirements covers audio length, file size, URL, and upload rules.
- Supported File Types lists the containers and codecs.
- Supported Languages lists the codes for the
languageparameter. - Errors explains each error message and how to fix it, including Upload errors and failures after a job is accepted.
- Changelog lists the releases.