Ask Anything

Ask Anything records one spoken question, transcribes it, sends that single prompt to your configured model, and shows the answer in a focused result panel.

How the flow works

  1. Start Ask Anything from the configured shortcut or menu action.
  2. Speak one focused question. The feature is designed for one-off answers, not long chat history.
  3. Stop recording. OpenTypeless sends the captured audio to the configured STT path.
  4. The transcript becomes the prompt for the configured LLM.
  5. Read the final answer in the result panel, then copy or return to your current app.

Credit and provider behavior

When Ask Anything uses hosted cloud STT or hosted LLM calls, that usage should count like other cloud features. BYOK or local provider paths depend on the user's own provider setup, so check pricing and quota before relying on heavy usage.

Frequently Asked Questions

Does Ask Anything keep chat context?

No. It is designed for one-off voice questions where the spoken transcript is enough context.

Does it need a text input box?

No. The microphone recording is the input. The result panel should show the final answer without requiring another send action.

Does Ask Anything use my quota?

Hosted STT and hosted LLM usage should count toward cloud entitlement. BYOK and local paths depend on your own provider billing and configuration.

When should I use it instead of normal dictation?

Use Ask Anything when you want an answer first. Use normal dictation when you want polished text inserted into the current app.

Related guides