Migrating
Migrating from Google Cloud Speech-to-Text
Config mapping, credential model differences, and feature parity.
Key differences
- GCP uses service account JSON credentials. Shunyalabs uses a simple Bearer API key — no credential file needed.
- GCP requires specifying
encodingandsampleRateHertzexplicitly. Shunyalabs auto-detects from the file. - GCP long-running jobs use an async polling model. Shunyalabs returns synchronously for files up to 15 minutes.
Config mapping
| GCP Parameter | Shunyalabs Equivalent | Notes |
|---|---|---|
| languageCode: "hi-IN" | language_code="hi" | Use 2-letter ISO code. |
| model: "latest_long" | model="zero-indic" | |
| enableSpeakerDiarization | enable_diarization=True | |
| enableWordTimeOffsets | result.segments | Segment-level only. |