• multimodal — A model that works with more than text — reading images, audio, or video, and sometimes generating them too.

Sign-language input is now a shipped phone keyboard modality, and the split design — pose extraction on device, translation on the server — is a reusable pattern for privacy-sensitive multimodal features.

• articlePutting sign language AI into users’ handsdeepmind.google
• articleIntelligent transcription with Gemini 3.5 Transcribedeepmind.google
• articleGemini 3.5 Transcribe now available on AI GatewayKevin Dawkins

Checking sign-in…

Loading comments…