SpeechBrain
Build speech recognition, synthesis and audio processing pipelines in PyTorch.
What’s included for free?
The toolkit is ongoing-free under Apache-2.0 with no provider minute quota; compute and pretrained models are chosen separately.
Where it runs: Self-hosted
Provider information ↗Checked 2026-11-01
Before you use it
Apache-2.0 covers the toolkit. Individual models, datasets and voices can use different licences.
Provider terms can change. Check current pricing and usage rights before starting a project.