Voice prompts — CodeAgent Mobile Wiki
The CodeAgent Mobile Wiki

Every feature, explained — organized by the path you actually take: set up, launch a session, supervise the run, ship the PR.

Free

Voice prompts

Send prompts to your agent by speaking instead of typing.

01 · COMPOSERRecording
Tap the mic
  • Microphone permission requested once
02 · SPEAKStreaming
Live transcript
  • Speech streams into the input as you talk
03 · SENDReady
Prompt ready
  • Edit the text like any typed message

Typing a long prompt on a phone keyboard is slow. Tap the mic icon in the composer and speak instead — speech-to-text streams your words into the input live, so you can watch the transcript build as you talk. Tap again to stop, edit if anything needs fixing, then send it like any other message.

Tap to start, tap to stop

One tap begins recording; tap again (or pause) to end it — no separate recording screen to navigate.

Transcript streams live

Recognized words land in the composer as you speak, not after you finish.

Pairs with Smart Composer

A voice transcript can feed straight into Smart Composer, which structures it into a cleaner prompt before you send.

First one is free

Every FREE user gets one voice prompt on the house — PRO unlocks unlimited voice transcription.

  1. 1
    Open the composer

    Start typing a message in any session, or go straight to the mic.

  2. 2
    Tap the mic icon

    Grant microphone access the first time — CodeAgent only asks once.

  3. 3
    Speak your prompt

    The transcript streams into the input live as you talk.

  4. 4
    Edit & send

    Fix anything the recognizer missed, then send exactly like a typed message.

NOTE

Voice prompts are plain speech-to-text — no AI restructuring happens here. For that, pair it with Smart Composer, which takes the raw transcript and helps you turn it into a better-structured prompt.

FAQ

Related

Unified Voice CommandSmart Composer@codeagent mentions

Ready to try it?

Start freeSee pricing