Create a note from an audio file using a two-step process. First, generate a presigned URL for audio upload, then create the note from the uploaded audio.
This feature is available for Enterprise plans only. Both endpoints require Enterprise-level API keys.

​
Step 1: Generate Upload URL

POST /v1/notes/from-audio/generate-upload-url

Generate a presigned URL for uploading an audio file to our servers.

​
Request Body

​
fileExtension
string
required
File extension of the audio file. Options: m4a, mp3, wav, aac
​
duration
number
required
Duration of the audio file in seconds (maximum 4 hours / 14,400 seconds)

​
Response

​
presignedUrl
string
The presigned URL for uploading the audio file to S3. This URL expires in 10 minutes.
​
audioUrl
string
The final URL where the audio file will be accessible after upload. Use this URL in Step 2.

​
Example Request

curl -X POST 'https://api.caret.so/v1/notes/from-audio/generate-upload-url' \
  -H 'Authorization: Bearer {api_key}' \
  -H 'Content-Type: application/json' \
  -d '{
    "fileExtension": "m4a",
    "duration": 1800
  }'

​
Example Response

{
  "presignedUrl": "https://caret-cdn.s3.ap-northeast-2.amazonaws.com/temp/note_upload_20241225_143052_abc123.m4a?X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=...",
  "audioUrl": "https://caret-cdn.s3.ap-northeast-2.amazonaws.com/temp/note_upload_20241225_143052_abc123.m4a"
}

​
Step 2: Upload Audio File

Use the presignedUrl from Step 1 to upload your audio file directly to our servers.

​
Example Upload

curl -X PUT '{presignedUrl}' \
  -H 'Content-Type: audio/m4a' \
  --data-binary '@/path/to/your/audio.m4a'

​
Step 3: Create Note from Audio

POST /v1/notes/from-audio

Create a note from the uploaded audio file using the audioUrl from Step 1.

​
Request Body

​
audioUrl
string
required
The audio URL returned from Step 1 (generate upload URL)
​
title
string
Optional title for the note. If not provided, a title will be generated automatically.
​
tags
array
Array of tag IDs to associate with the note
​
kind
string
Type of meeting. Options: online, in-person, podcast. Default: in-person
​
existingNoteId
string
Optional existing note ID to update instead of creating a new note

​
Response

​
noteId
string
The ID of the created note

​
Example Request

curl -X POST 'https://api.caret.so/v1/notes/from-audio' \
  -H 'Authorization: Bearer {api_key}' \
  -H 'Content-Type: application/json' \
  -d '{
    "audioUrl": "https://caret-cdn.s3.ap-northeast-2.amazonaws.com/temp/note_upload_20241225_143052_abc123.m4a",
    "title": "Weekly Team Meeting",
    "tags": ["01887270-23f5-7da0-b95c-9a9e9ebc3c25"],
    "kind": "online"
  }'

​
Example Response

{
  "noteId": "01887270-12d4-7da0-b95c-9a9e9ebc3b13"
}

​
Complete Workflow Example

Here’s a complete example of the audio-to-note conversion process: 
# Step 1: Generate upload URL
UPLOAD_RESPONSE=$(curl -X POST 'https://api.caret.so/v1/notes/from-audio/generate-upload-url' \
  -H 'Authorization: Bearer {api_key}' \
  -H 'Content-Type: application/json' \
  -d '{
    "fileExtension": "m4a",
    "duration": 1800
  }')

# Extract URLs from response
PRESIGNED_URL=$(echo $UPLOAD_RESPONSE | jq -r '.presignedUrl')
AUDIO_URL=$(echo $UPLOAD_RESPONSE | jq -r '.audioUrl')

# Step 2: Upload audio file
curl -X PUT "$PRESIGNED_URL" \
  -H 'Content-Type: audio/m4a' \
  --data-binary '@/path/to/your/audio.m4a'

# Step 3: Create note from audio
curl -X POST 'https://api.caret.so/v1/notes/from-audio' \
  -H 'Authorization: Bearer {api_key}' \
  -H 'Content-Type: application/json' \
  -d "{
    \"audioUrl\": \"$AUDIO_URL\",
    \"title\": \"Weekly Team Meeting\",
    \"kind\": \"online\"
  }"

​
Processing Status

After creating a note from audio, the note will go through several processing stages:
  1. transcript-processing: The audio is being transcribed
  2. summarizing: The transcript is being summarized and processed
  3. completed: The note is fully processed and ready
You can check the processing status by getting the note details using the returned noteId.

​
Important Notes

  • Audio files must be under 4 hours in duration
  • Supported formats: M4A, MP3, WAV, AAC
  • The presigned URL expires in 10 minutes after generation
  • Processing time depends on audio length and quality
  • Enterprise plan required for this feature