Translate audio
Feed us a recording in one language and get back text in another, without a separate transcription step in between. It is built for the everyday case of content that was spoken in a language your audience does not read.
Run it online
Run this on our servers with your account. Free tools run in your browser; this one bills your KIT balance per the price above.
The problem it removes
Translating spoken content traditionally means two jobs stitched together: transcribe, then translate, each with its own settings, its own failure modes, its own bill. Errors compound — a mangled transcript produces a mangled translation no matter how good the translation step is. audio.translate treats it as one job, so the model doing the translating has the audio context, not just a flawed intermediate transcript.
What you send and what comes back
POST /audio/translate with the source audio and, if you know it, the target language; the call returns a task_id right away while the work runs asynchronously. You get delivery through your webhook or a signed link valid for 24 hours, with the translated text structured for direct use — captions, articles, support tickets, whatever your pipeline expects next.
Speech translation is older than it looks
Machine speech translation has roots going back to research systems from the 1980s and 90s that chained recognition, translation and synthesis together with visible seams. Cascaded and end-to-end approaches have both matured since, and the practical result for you is translation quality that holds up on real conversational audio, not just clean studio narration.
Where accuracy naturally varies
Overlapping speakers, heavy accents, idiom and industry jargon all affect any speech translation system, ours included; we do not claim figures we cannot stand behind, and we recommend a light human pass for anything published under your name. What we do guarantee is that a task which fails outright — corrupted audio, unsupported format after three retries — is never billed.
Fitting it into a pipeline
Chain audio.detect_language ahead of it when the source language is unknown, or call it directly when you already know what you are working with. Content teams use it to open a single foreign-language interview into a target-language draft in one call instead of running transcription and translation as separate jobs with separate costs.
What you can do with it
Foreign interview to publishable draft
A newsroom translates a source's spoken interview directly into a draft article in the newsroom's working language.
Customer call review
A support team translates recorded calls from international customers so managers who don't speak the caller's language can review quality.
Cross-border research interviews
A research team translates field interviews recorded abroad into their working language for coding and analysis.
Multilingual training footage
An L&D team translates a training recorded by a regional office into the company's other operating languages for internal distribution.
FAQ
How does the audio translation API work?
Send audio to POST /audio/translate, get a task_id back immediately, and receive the translated text by webhook or a signed link valid for 24 hours once processing completes.
Do I need to specify the source language?
It helps but is not always required; if you don't know it, run audio.detect_language first and pass the result in.
Is audio translation free?
No, there is no free tier or trial. It costs $0.004 per request plus $0.0185 per minute, and a task that fails after retries is never charged.
Does it transcribe and translate in one call?
Yes, transcription and translation happen as a single task so you get target-language text directly, without managing two separate jobs.
How accurate is the translation?
Accuracy depends heavily on audio quality, accents and overlapping speech; we don't publish invented accuracy figures, and we recommend a human review pass for published or legal use.
Can it handle long recordings?
Yes, it processes asynchronously specifically so longer files — meetings, interviews, lectures — don't time out a request; you get notified when the result is ready.
What happens if the task fails?
We retry automatically up to three times; if it still fails you get a clear error and are never charged for that task.
Is my audio stored or used for training?
No. Audio and results are deleted after the retention period and are never used to train models.
For developers — API access
Everything on this page is available programmatically. This section is for teams who want to wire it into their own systems; everyone else can just use the tool above.
API endpoint
Prefer to automate it? One authenticated POST creates the task; the result comes back by webhook or a signed link. The same capability also runs here on the web, and soon from our app, email and Telegram.
Call it from your stack
curl -X POST https://api.kit.forhosting.com/audio/translate \
-H "Authorization: Bearer $KIT_KEY" \
-H "Content-Type: application/json" \
-d '{"audio":"https://ejemplo.com/audio.mp3"}'const res = await fetch("https://api.kit.forhosting.com/audio/translate", {
method: "POST",
headers: {
"Authorization": `Bearer ${process.env.KIT_KEY}`,
"Content-Type": "application/json"
},
body: JSON.stringify({
"audio": "https://ejemplo.com/audio.mp3"
})
});
const { task_id } = await res.json();import os, requests
res = requests.post(
"https://api.kit.forhosting.com/audio/translate",
headers={"Authorization": f"Bearer {os.environ['KIT_KEY']}"},
json={
"audio": "https://ejemplo.com/audio.mp3"
},
)
task_id = res.json()["task_id"]<?php
$res = file_get_contents("https://api.kit.forhosting.com/audio/translate", false, stream_context_create([
"http" => [
"method" => "POST",
"header" => "Authorization: Bearer " . getenv("KIT_KEY") . "\r\nContent-Type: application/json",
"content" => '{"audio":"https://ejemplo.com/audio.mp3"}',
],
]));
$task = json_decode($res, true);body := bytes.NewBufferString(`{"audio":"https://ejemplo.com/audio.mp3"}`)
req, _ := http.NewRequest("POST", "https://api.kit.forhosting.com/audio/translate", body)
req.Header.Set("Authorization", "Bearer "+os.Getenv("KIT_KEY"))
req.Header.Set("Content-Type", "application/json")
res, _ := http.DefaultClient.Do(req)Example request
{
"audio": "https://ejemplo.com/audio.mp3"
}Example response
{
"task_id": "tsk_a1b2c3d4e5f6a1b2c3d4e5f6",
"type": "audio.translate",
"status": "queued",
"_links": {
"result": "/tasks/tsk_…/result"
}
}The API is asynchronous: the call returns a task_id immediately and the result arrives by webhook. Polling is capped at 1 req/s per task.
Pricing
Published price — no tokens, no invented credits. A failed task is never charged.
Limits
max_mb | 200 |
max_minutes | 180 |
Errors
| HTTP | Code | Meaning |
|---|---|---|
401 | unauthorized | Missing or invalid API key. |
402 | insufficient_balance | Your balance doesn't cover the task price. |
404 | unknown_type | That task type doesn't exist. |
429 | rate_limited | Too many requests. Use the webhook instead of polling. |
422 | task_failed | The task failed after 3 retries. You are never charged for it. |