Speech and audio models

Yes. Products in speech and audio models are growing faster than their peers in agent skills, most of all in agent skills (1.9× the typical product).

MERaLiON-3-3B-ASR (21×, open models), voiceclap-large-v2 (18×, open models) and speakoflow-mini (9.2×, open models).

Agent skills: 1.86× the typical product, across 8 products. Open models: 1.04× the typical product, across 787 products. Workplace add-ons: 1.03× the typical product, across 8 products. Code editor extensions: 1.02× the typical product, across 16 products.

It ranks #11 of 152 themes by how much faster its products grow than their peers.

Models for speech recognition, text-to-speech, voice cloning, and audio generation Vela Breakout groups products into themes by what they do, read from each product's own description.