Crane NeMo ASR 1.2

Crane AI Labs

Luganda and Swahili transcription.

Pilot betaInvited workspaces · Cape Town, South Africa

What you can build.

Capabilities available in the Embiro pilot.

InputAudio

OutputText

Luganda transcription
Swahili transcription
English transcription

Know the limits.

Languages: lg · Luganda, sw · Swahili, en · English
Audio format: Mono PCM16 WAV at 16 kHz, up to 60 seconds
Processing: Offline transcription; review accuracy, especially Luganda

Full pilot service notes

Luganda, Swahili and English beta; mono PCM16 WAV, 16 kHz, up to 60 seconds; offline transcription. Rounded up to whole seconds. Review accuracy, especially Luganda.

Requests share your workspace’s allowance, rate limit and concurrency limit. Compute runs in scheduled sessions; model access does not guarantee immediate capacity.

Make your first request.

Use a server-side workspace API key from Platform → API keys. Examples use https://api.embiro.ai/v1 and the exact pilot model ID.

API model IDcrane-nemo-asr-1.2

POST/v1/audio/transcriptions

Prepare clip.wav as mono PCM16 WAV at 16 kHz, up to 60 seconds. This example uses Swahili (sw); change language to lg or en for Luganda or English.

# Set EMBIRO_API_KEY in your shell first.
: "${EMBIRO_API_KEY:?Set your workspace API key first}"
curl --fail-with-body --silent --show-error "https://api.embiro.ai/v1/audio/transcriptions" \
  -H "Authorization: Bearer $EMBIRO_API_KEY" \
  -F "model=crane-nemo-asr-1.2" \
  -F "language=sw" \
  -F "response_format=json" \
  -F "file=@clip.wav;type=audio/wav"

What comes back

A transcript in text. Silent audio may return an empty transcript.

Cold starts, errors and safe retries

Use the Queued job example when a request can wait for compute. Save its ID and poll that job.
For 401/403, check the key and its model permissions. For 400, check the request and format limits.
Respect Retry-After on 429/503. A timeout after dispatch can be ambiguous; check request logs before submitting again.
These examples do not automatically retry paid work. A job retry must reuse the saved idempotency key and identical payload.

Open API reference

Under the hood.

Publisher specifications. These do not override the Embiro service limits above.

Published context: Not specified
Parameters: 0.638B
Model size: Light

Luganda, Swahili and English beta. Offline WAV transcription up to 60 seconds; review accuracy before use.

Read the publisher’s model card

No verified benchmark scores are listed in this catalogue for this model.