sdk.intelligence). It covers on-device inference through the selected backend, registered cloud backends, runtime diagnostics, and active persona state. Common operations include chat completions (streaming or one-shot), text generation, backend listing, cloud key registration, and state inspection.
Apps access it through the platform bridge object:
intelligence in your shell manifest under requires.managers.
Full reference
The complete intelligence facade reference — method signatures, parameter tables, event payloads, and return types — lives in the in-app Developer panel.Open the Developer panel
- Open Settings → Experimental and toggle Developer Mode on.
- A terminal icon appears in the top right of the app. Click it to open the Developer panel.
- Go to the Documentation section → Docs tab →
intelligence.
Learn more
Build apps
Build your first app using the webAI SDK and platform facades.
On-device AI
How the AI runtime works — backends, model tiers, LoRA adapters.