Skip to content

Your preferences

Everything under Preferences is about you, not any one project. Open it from the account menu in the sidebar (your username, then Preferences), or go to aquilla.app/preferences. The page’s own words set the scope: “Personal preferences that apply to you across all projects on this device.” That last part matters: most of these settings live in your browser, so they follow you per device, not per account.

The Preferences page with a General section (Theme, UI language, Share usage data, Workspace), an AI & personalization section (Translator profile, AI provider keys, Local models), and an Account section (Usage, API tokens)
Three groups: how the app looks and behaves, what the AI knows about you, and your account's usage and tokens.
  • Theme: follow your system appearance, or pin Light or Dark on this device.
  • UI language: “The language the app’s own interface (menus, buttons, messages) is shown in.” This is separate from any project’s languages and from the AI’s reply language.
  • Share usage data: product analytics, on or off. The description is explicit about the boundary: “Events like project creation, exports, and AI translations. Never the contents of your translations or files.”
  • Workspace: where the editor sidebar’s tab strip lives (Left rail or Top bar), and whether replacing an existing translation asks for confirmation first.

Translator profile is the interesting one: a short, entirely optional form the AI reads so “its summaries and answers fit your context — and so it replies in your language.” Assistant language, age, education, translation experience, geographical setting, and anything else you want it to weigh. All of it stays on your device.

The Translator profile form with optional fields for assistant language, age, gender, education, religious background, translation experience, and geographical setting, filled in with Spanish, first project, and an urban Latin America setting
A few optional facts about you, and the AI's answers and summaries start fitting your situation.

AI provider keys holds your personal key override: a key saved here is used for your own AI calls across projects on this browser, on top of whatever each project configures. Local models manages the on-device models (Whisper transcription and the local voices), which download once and are shared by every project on the device.

Usage shows your recent AI and audio activity: requests today, audio generated today, and a 7-day audio history.

API tokens mints personal access tokens for the Agent API, so an external agent or script can act with your access. Each token gets a name, a scope (an organization, optionally narrowed to one project), an expiry, and a mode:

  • Ask — every write waits for your approval.
  • Act — writes apply immediately. Requires a project below where you’re a maintainer.

The token’s plaintext is shown once at minting; after that it can only be revoked. Treat it like a password, because the page says exactly that. Approving agent changes walks through what a connected agent’s requests look like and how the Ask mode’s approvals work.