Back to Support
Guide

How to use voice features

Use the microphone button to speak your questions, the speaker button to hear answers read aloud, and voice commands to manage goals, budgets, and transactions hands-free. All voice features are available to every user at no extra cost.

  1. 1

    Navigate to Dashboard > Finance > Ask (or click the "Ask" tab).

  2. 2

    Click the microphone button in the chat input area, or hold the keyboard shortcut, to start recording.

  3. 3

    Speak your question. Recording stops automatically after 3 seconds of silence, or click the mic button again to stop manually.

  4. 4

    Review the transcript in the input field — edit if needed, then press Send.

  5. 5

    Hover over the assistant's response and click the speaker button to hear it read aloud.

Voice Input (Speech-to-Text)

The microphone button sits in the chat input area, next to the text field. Tap it to start recording:

  • Recording indicator — A pulsing red ring appears while recording is active.
  • Transcription — When you stop recording, a spinner appears while your speech is transcribed. The resulting text is placed in the input field so you can review or edit before sending.
  • Voice Activity Detection — Recording stops automatically after 3 seconds of silence, so you don't need to manually hit stop.
  • Limits — Each recording must be at least 1 second and no longer than 2 minutes. Voice input is rate-limited to 20 requests per minute.
  • Browser support — Your browser must support the MediaRecorder API. Chrome and Edge have full support. Firefox and Safari (14.1+) have partial support. The mic button is automatically hidden on unsupported browsers.

For general guidance on using the Ask AI chat interface, see How to Use Ask AI.

Listening to Responses (Text-to-Speech)

Hover over any assistant message to reveal the speaker button in the top-right corner of the message bubble.

  • Play/Stop — Click the speaker icon to start playback. Click it again (or start playing a different message) to stop.
  • Single player — Only one message plays at a time. Starting a new message automatically stops the previous one.
  • Caching — Replaying the same message reuses cached audio — no additional API calls or credits are used.
  • Clean audio — Markdown formatting (bold, headers, links, code) is stripped before synthesis so the audio sounds natural.

Voice-Driven Financial Actions

You can execute real financial operations entirely through voice. Just speak naturally — Claude automatically detects whether you are asking a question or requesting an action.

Supported actions include:

  • Goals — Create, update, or delete savings and debt goals (e.g., "Create a vacation savings goal for $3,000 by December")
  • Budgets — Create budgets and add or update budget categories (e.g., "Set my dining budget to $400 this month")
  • Transactions — Update transaction category, merchant, or notes (e.g., "Move my Starbucks transaction to the dining category")
  • Recurring streams — Update amount, merchant, frequency, or category for recurring transactions

Safety for destructive actions: Delete operations always require a two-turn confirmation. Claude will ask "Are you sure?" before executing any deletion.

Keyboard Shortcut

You can start and stop voice recording with a keyboard shortcut:

  • Default shortcut — Alt + Shift + M on Windows/Linux, or Cmd + Option + M on Mac
  • Hold-to-record — Hold the shortcut keys to record, release to stop
  • Customize — Go to Settings > Shortcuts to change the shortcut to any key combination you prefer
  • Priority — When the feedback modal is open, its voice button takes priority over the keyboard shortcut

Feedback Voice Button

The feedback modal (accessible from the header or sidebar) includes a separate voice recorder:

  • A larger blue circular button with an elapsed time timer
  • Same 2-minute maximum recording length
  • Speak your feedback instead of typing — your speech is transcribed and placed in the feedback text field

Compatibility and Requirements

BrowserSupport
Chrome (desktop & Android)Full
EdgeFull
FirefoxPartial
Safari 14.1+Partial
  • A working microphone is required for voice input
  • Voice features currently support English only
  • Available to all users, including the free tier

Tips for Using Voice

  • Speak naturally — Claude understands conversational phrasing, no special commands needed
  • Be specific with numbers and dates — "Save $500 per month toward my emergency fund by June" works better than "save some money"
  • Review the transcript — Glance at the transcribed text before sending to catch any errors
  • Try voice for complex questions — Speaking naturally often produces more detailed questions than typing
  • Check microphone permissions — If recording fails to start, make sure your browser has microphone permission enabled for abda

Does voice input cost extra?

No. Voice uses the same credit system as text. The first question in a new conversation costs 1 credit; all follow-ups in the same conversation are free.

Why did my recording stop automatically?

Voice Activity Detection stops recording after 3 seconds of silence. You may also have hit the 2-minute maximum. Tap the mic button to start a new recording.

Can I use voice on my phone?

Yes. Voice input works on Chrome for Android. Safari on iOS has partial support starting from version 14.1.

Is my voice data stored?

Voice audio is transcribed and then discarded. Only the text transcript is sent to the AI and saved as part of your conversation.

What languages are supported?

Voice features currently support English only.