-
-
Notifications
You must be signed in to change notification settings - Fork 683
Managed services
The Managed speech recognition service enables you to immediately start using the extension, without having to sign up for third-party services.
The extension may use local and remote speech recognition services to solve a challenge, depending on how it was configured from the extension's options. When both local and remote services are enabled, the extension will first attept to solve the challenge using a local service.
The extension uses the Prompt API to access Gemini Nano, a built-in AI model in Chrome.
The model only works on modern devices which support the GPU backend.
Visit Settings > System in Chrome to enable On-device AI. You can learn more
about the current state of the installed model in chrome://on-device-internals.
Chrome automatically downloads the model on first use. The status of the model
in the extension's options must be ready to use to solve challenges
using the local service.
System requirements:
- Browser: Chrome 148
- Operating system: Windows 10 or 11, macOS 13+, Linux, or ChromeOS (from Platform 16389.0.0 and onwards) on Chromebook Plus devices
- Storage: at least 22 GB of free space on the volume that contains your Chrome profile
- GPU: more than 4 GB of VRAM
On resource-constrained devices a lightweight version of Whisper is used.
Visit the extension's options to download and store the 50 MB model
in the browser cache. The status of the model must be ready to use to solve
challenges using the local service.
The extension uses Wit.ai with API keys set up by project maintainers.