Transcribe audio

curl --request POST \
  --url https://api.trymorse.ai/v1/audio/transcriptions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form model=whisper-large-v3-tee \
  --form file=@example-file \
  --form language=en

import requests

url = "https://api.trymorse.ai/v1/audio/transcriptions"

files = { "file": ("example-file", open("example-file", "rb")) }
payload = {
    "model": "whisper-large-v3-tee",
    "language": "en"
}
headers = {"Authorization": "Bearer <token>"}

response = requests.post(url, data=payload, files=files, headers=headers)

print(response.text)

const form = new FormData();
form.append('model', 'whisper-large-v3-tee');
form.append('file', '(the recording, as a 16-bit PCM WAV file)');
form.append('language', 'en');

const options = {method: 'POST', headers: {Authorization: 'Bearer <token>'}};

options.body = form;

fetch('https://api.trymorse.ai/v1/audio/transcriptions', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "text": "And so, my fellow Americans, ask not what your country can do for you.",
  "usage": {
    "type": "duration",
    "seconds": 11
  }
}

{
  "error": {
    "message": "file must be a WAV file of 16-bit PCM, mono or stereo, at 8 to 48 kHz.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_parameter"
  }
}

{
  "error": {
    "message": "Incorrect API key provided.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_api_key"
  }
}

{
  "error": {
    "message": "The account balance does not cover this request. Add credit.",
    "type": "insufficient_balance",
    "param": null,
    "code": "insufficient_balance"
  }
}

{
  "error": {
    "message": "This account has been disabled. Contact support.",
    "type": "invalid_request_error",
    "param": null,
    "code": "account_disabled"
  }
}

{
  "error": {
    "message": "The requested model does not exist.",
    "type": "invalid_request_error",
    "param": null,
    "code": "model_not_found"
  }
}

{
  "error": {
    "message": "The request body did not arrive in time.",
    "type": "invalid_request_error",
    "param": null,
    "code": "request_timeout"
  }
}

{
  "error": {
    "message": "Request body is too large.",
    "type": "invalid_request_error",
    "param": null,
    "code": "request_too_large"
  }
}

{
  "error": {
    "message": "Rate limit reached for this account. Retry after the time in the Retry-After header.",
    "type": "requests",
    "param": null,
    "code": "rate_limit_exceeded"
  }
}

{
  "error": {
    "message": "The transcription provider failed.",
    "type": "upstream_error",
    "param": null,
    "code": "upstream_error"
  }
}

{
  "error": {
    "message": "The transcription provider is busy. Try again shortly.",
    "type": "upstream_error",
    "param": null,
    "code": "upstream_busy"
  }
}

API Reference

Transcribe audio

Turns a recording of speech into text, with OpenAI’s transcription request and reply. Send the recording as a WAV file of 16-bit PCM in a multipart form, at most 5 minutes long and 10 MiB. Each call costs a flat 200 micros, a fiftieth of a credit. A failed call costs nothing. whisper-large-v3-tee runs Whisper Large v3 on NEAR’s attested hardware, and Morse sends the audio to it only over the TLS key its attestation report binds.

Transcription is not offered to API keys for now: GET /v1/models lists no such model to a key, and a key’s request here gets 404 model_not_found. A Morse app login can still use it.

POST

audio

transcriptions

Transcribe audio

curl --request POST \
  --url https://api.trymorse.ai/v1/audio/transcriptions \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: multipart/form-data' \
  --form model=whisper-large-v3-tee \
  --form file=@example-file \
  --form language=en

import requests

url = "https://api.trymorse.ai/v1/audio/transcriptions"

files = { "file": ("example-file", open("example-file", "rb")) }
payload = {
    "model": "whisper-large-v3-tee",
    "language": "en"
}
headers = {"Authorization": "Bearer <token>"}

response = requests.post(url, data=payload, files=files, headers=headers)

print(response.text)

const form = new FormData();
form.append('model', 'whisper-large-v3-tee');
form.append('file', '(the recording, as a 16-bit PCM WAV file)');
form.append('language', 'en');

const options = {method: 'POST', headers: {Authorization: 'Bearer <token>'}};

options.body = form;

fetch('https://api.trymorse.ai/v1/audio/transcriptions', options)
  .then(res => res.json())
  .then(res => console.log(res))
  .catch(err => console.error(err));

{
  "text": "And so, my fellow Americans, ask not what your country can do for you.",
  "usage": {
    "type": "duration",
    "seconds": 11
  }
}

{
  "error": {
    "message": "file must be a WAV file of 16-bit PCM, mono or stereo, at 8 to 48 kHz.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_parameter"
  }
}

{
  "error": {
    "message": "Incorrect API key provided.",
    "type": "invalid_request_error",
    "param": null,
    "code": "invalid_api_key"
  }
}

{
  "error": {
    "message": "The account balance does not cover this request. Add credit.",
    "type": "insufficient_balance",
    "param": null,
    "code": "insufficient_balance"
  }
}

{
  "error": {
    "message": "This account has been disabled. Contact support.",
    "type": "invalid_request_error",
    "param": null,
    "code": "account_disabled"
  }
}

{
  "error": {
    "message": "The requested model does not exist.",
    "type": "invalid_request_error",
    "param": null,
    "code": "model_not_found"
  }
}

{
  "error": {
    "message": "The request body did not arrive in time.",
    "type": "invalid_request_error",
    "param": null,
    "code": "request_timeout"
  }
}

{
  "error": {
    "message": "Request body is too large.",
    "type": "invalid_request_error",
    "param": null,
    "code": "request_too_large"
  }
}

{
  "error": {
    "message": "Rate limit reached for this account. Retry after the time in the Retry-After header.",
    "type": "requests",
    "param": null,
    "code": "rate_limit_exceeded"
  }
}

{
  "error": {
    "message": "The transcription provider failed.",
    "type": "upstream_error",
    "param": null,
    "code": "upstream_error"
  }
}

{
  "error": {
    "message": "The transcription provider is busy. Try again shortly.",
    "type": "upstream_error",
    "param": null,
    "code": "upstream_busy"
  }
}

Authorizations

Authorization

string

header

required

An API key from the Morse app. It starts with sk_morse_.

Body

multipart/form-data

model

string

required

A transcription model's id, such as whisper-large-v3-tee.

file

required

The recording, as a WAV file of 16-bit PCM, mono or stereo, at 8 to 48 kHz, read from its bytes whatever its file name or Content-Type. At most 5 minutes long, and 10 MiB with the other fields.

language

string

The spoken language as an ISO-639-1 code Whisper transcribes, such as en. Without it, Whisper detects the language.

response_format

any

Only json, the default.

Response

The transcript.

text

string

required

The transcript, without surrounding whitespace.

usage

object

required

Show child attributes