Single-surface assistants are clever in a vacuum. They can summarize an email or draft a meeting brief, but they don't know what else is happening in your life. Aware's assistant is built to reason across surfaces, not within one.
A request like 'reschedule my Thursday lunch to next week, push the dependent invoice payment too, and tell Sarah on Slack' touches three apps. A unified assistant handles it as one action.
Aware indexes events from each connected surface into a shared memory. The assistant queries that memory before generating any reply, which is why it can answer with specifics your other tools can't.
Aware routes to the best available model per task — Gemini, GPT, and Claude, with automatic fallback.
Yes. Long-term memory survives sessions and is editable.
Yes — every connection is toggleable and you can pause memory at any time.
Plugins are tools the model calls. Aware's memory is the model's working context — always on, not requested.