Official documentation
Live transcription with Parakeet Unified (Petal documentation)
Aayush9029, GitHub (Aayush9029/petal), 2026
Read the originalRead on October 25, 2026
- What it is
- A documentation page in the Petal repository about the Parakeet Unified 0.6B model for live transcription. It says the model runs locally through FluidAudio and CoreML and that Petal downloads about 625 MB of model files.
- How much weight it carries
- The maker's own developer documentation. It describes one selectable model and is not a statement about the app's default. It does not say whether the other speech models run on the device.
Facts taken from this source
- Petal's documentation says its Parakeet Unified model runs locally.
“The model runs locally through FluidAudio and CoreML.” (Live transcription with Parakeet Unified)