§ Enhanced mode
Enhanced mode
Enhanced runs Polish, then attaches the focused app and (opt-in) the active selection and a window screenshot, then routes the transcript through a model rewrite. The local MLX polish model is the default; OpenAI / Anthropic / OpenAI-compatible endpoints are opt-in BYOK.
Who this is for
Users who want app-aware polish — different rewrites for Slack vs email vs a code review comment — and users who lean on translation, context-grounded vocabulary, or automation pipelines tied to dictation.
What "context-aware" actually means
Enhanced reads three signals before it calls a model: the frontmost app and its bundle ID, the active selection (if Accessibility is granted), and a single window screenshot (only when the per-app screenshot toggle is on for that app class).
These signals are bounded — there is a per-context token budget and a payload metric in Settings. Dettivo will surface a status pill (Ready / Limited / Blocked) in the menu bar and the completion notice so you know what context made it into the run.
Context toggles
Settings › AI Processing › App context
- Enable app context
- Master switch. With it on, frontmost app + bundle ID are attached to every Enhanced run.
- Use frontmost window screenshots
- Optional. Requires Screen & System Audio Recording permission. Bounded to a single frame, only when you enable it for an app class.
- Code apps · Mail apps · Browsers · Other apps
- Per-class screenshot toggle. Most users enable code apps only.
- Remote endpoint trust
- If you point Enhanced at an OpenAI-compatible endpoint, this row controls whether Dettivo allows that exact base URL.
Model choices for Enhanced
Launching April 2026
Be first on the launch list.
Dettivo opens for new purchases in April 2026. One-time license, three Macs, lifetime updates.