Quickstart¶
Get transcribing in 5 minutes.
Prerequisites¶
- An Icana API key (see Authentication)
- An audio file to transcribe
Step 1: Upload Your Audio File¶
Upload the audio file to get an S3 URI:
Response:
{
"message": "File uploaded successfully",
"s3_uri": "s3://your-upload-bucket/unique-id/meeting.mp3",
"original_filename": "meeting.mp3",
"job_id_suggestion": "unique-id"
}
Save the s3_uri from the response.
Step 2: Submit Transcription Job¶
Submit the uploaded file for transcription and diarisation:
curl -X POST https://api.icana.ai/transcribe \
-H "X-API-Key: your-api-key" \
-H "Content-Type: application/json" \
-d '{
"s3_uri": "s3://your-upload-bucket/unique-id/meeting.mp3",
"language": "en"
}'
Response:
{
"message": "Batch job submitted successfully",
"aws_batch_job_id": "abc-123-def-456",
"batch_job_name": "whisper-pyannote-job-unique-id"
}
Save the aws_batch_job_id.
Step 3: Poll for Results¶
Check the job status until it completes:
When complete, the response includes the transcription and diarisation:
{
"aws_batch_job_id": "abc-123-def-456",
"status": "COMPLETE",
"message": "Batch job completed successfully.",
"transcription": "G'day everyone, thanks for joining today's standup...",
"diarization": "[ 0m0s - 0m5s ] Speaker 1: G'day everyone, thanks for joining today's standup."
}
Tip
While waiting, you can check the overall queue status with GET /queue/status to see how many jobs are in the pipeline.
Step 4: Clean Up (Optional)¶
Delete the uploaded and output files when you no longer need them:
curl -X DELETE https://api.icana.ai/delete/abc-123-def-456 \
-H "X-API-Key: your-api-key" \
-H "Content-Type: application/json" \
-d '{
"s3_uri": "s3://your-upload-bucket/unique-id/meeting.mp3"
}'
Next Steps¶
- Learn about authentication
- Explore all API endpoints
- See error handling
- Use the Python SDK or JavaScript SDK for a simpler interface