Voice input for AI agents and code editors
Dictate prompts to Claude Code, Cursor, Copilot or any terminal agent, and the prose around your code in VS Code: commit messages, PR descriptions, docstrings. The words land in the input box, not in the agent, including over SSH.
Free tier: 30 minutes a month, no card.
Can you dictate prompts to an AI agent?
Yes, with dictation that types rather than sends. VibeVoice puts the transcript into whichever text field has focus — a terminal running an agent, a browser chat box, an editor panel — and stops there. You read it, fix a term, add what you forgot, and send when you mean to.
What each agent’s own documentation says
Every cell is from the vendor’s own documentation or repository; Claude Code’s is the most thorough.[1] Two of the seven do editable dictation natively and are here for that reason.
| Agent | Native voice | What its docs describe |
|---|---|---|
| Copilot CLI | Yes | Inserts the transcript at the cursor to edit before submitting. Needs a runtime and model download, and lists English and Spanish.[2] |
| Aider | Yes | /voice records inside the chat and writes the speech as if typed — within Aider’s own session.[3] |
| Claude Code | With limits | Local mic only, not over SSH or remote VS Code; in tap mode, 15s silence and a two-minute cap, and it auto-submits.[1] |
| Codex | In the app | Voice lives in the ChatGPT desktop app and its session model: a chat must start in voice mode, one at a time, on an eligible plan.[4] |
| Gemini CLI | Proposed | An open request is answered with OS dictation or a third-party setup;[5] another asks for cursor-aware insertion because transcripts append to the end of the buffer.[6] |
| Kimi CLI | None | The repository documents terminal setup, slash commands and MCP tools, and no voice capture.[7] An open issue asks for it; the reply points at macOS Dictation.[8] |
| Pi | Packages | Third-party packages:[9] some run recognition on your machine,[10] one sends audio to a paid service,[11] and send behaviour differs by package.[12] |
A prompt is not a message
A wrong word in a message to a colleague is a typo and they work out what you meant. A wrong word in an instruction gets carried out.
In VS Code or Cursor: the prose around the code
The same hotkey works in the commit message box, a PR description, a docstring, a review comment, the chat panel and the integrated terminal. That is where most of a developer’s English goes, and it is what gets cut short when typing it costs more than the change did. Three fifty-word paragraphs of commit rationale take about 60 seconds to say.
It does not need to know which one you use
If it has a text field and it has focus, the words go in; the exception is programs that deliberately reject synthetic input. The row below is where people dictate today, not a list of supported agents.
What using it actually looks like
There is nothing to install into the agent or the editor, and nothing for either to support.
Put the cursor where you would type
The prompt line of a terminal agent, a chat box, an editor panel, the commit message field.
Press your hotkey and think out loud
Including the part where you change your mind. The text appears about half a second after you stop.
Edit it, then send it yourself
Fix a term, paste a stack trace under it. Nothing is sent until you press Enter.
What it does not do
- No integration, no plugin, no MCP server, and no sending. It types into the focused field, so it cannot read the agent’s output or act on it; if you want speak-and-go, a built-in voice mode is the right tool.
- It does not know your codebase. Common library names usually come through; internal service names and unusual identifiers can come back as ordinary words, and there is no custom vocabulary yet.
- Transcription happens on our servers, so it needs a connection. If prompts must never leave the machine, run something local instead.
Common questions
Does it work over SSH?+
Yes. The client runs on your own machine and types into the window in front of you, wherever the agent runs. Claude Code’s own voice dictation needs a local microphone and does not work in SSH or remote VS Code sessions, by its documentation.
Does dictating cost extra tokens?+
Not with this. The agent receives text, exactly as if you had typed it, and is billed as text. Speaking directly to a model that accepts audio is a different path with different billing, and you should check that model’s pricing rather than assume.
Is there a VS Code or Cursor extension?+
No, and none is needed. The client types into whichever field has focus at the operating-system level, so the editor, the terminal, the commit box and any AI chat panel all work without the editor knowing about it, and an editor update cannot break it.
Can I dictate code with it?+
You can, but it is slower than typing. Code is punctuation-dense and case-sensitive, and general dictation has no symbol grammar. Dictate the prompt or the explanation and type the code. If you need to write code by voice because a keyboard is not an option, you want a voice-coding tool built for that, with a spoken grammar you learn.
Sources
Checked 19 August 2026; Claude Code’s page re-read 28 October 2026. These products change without notice, so check the source before relying on a line of the table.
- Anthropic — Claude Code documentation, “Voice dictation”code.claude.com
- GitHub Docs — “Use voice input with Copilot CLI”docs.github.com
- Aider — “Voice-to-code with aider”aider.chat
- OpenAI — Codex documentation, “ChatGPT Voice”developers.openai.com
- gemini-cli issue #1982 — “Audio input instead of typing”github.com
- gemini-cli issue #25494 — voice transcriptions append to the end of the input buffergithub.com
- Moonshot AI — kimi-cli repository READMEgithub.com
- kimi-cli issue #1934 — “Feature Request: Voice Input Mode for Terminal and Web UI”github.com
- Pi — package cataloguepi.dev
- Pi — pi-voice-stt package pagepi.dev
- Pi — pi-vox package pagepi.dev
- Pi — picrophone package pagepi.dev
The same hotkey types into Slack, Gmail and Google Docs. Comparing dictation tools first? Start at the comparison hub.
Last updated .