Dictation in Zed on a Mac
Zed's Agent Panel is a natural place for voice. Prompts there are long, and Zed is fast enough that typing them feels like the slow part. Zed doesn't have dictation of its own yet, though, and some outside dictation tools have trouble with its agent input. Here's what Zed's docs and issue tracker say, and a setup that works.
What Zed has today
Zed's Agent Panel is where you work with AI agents that read files, edit code and run commands. You open it with agent: new thread from the Command Palette or the ✨ icon in the status bar, type in the message editor and press Enter to send. Typing @ adds context: files, folders, symbols, earlier threads, diagnostics, branch diffs and URLs. The panel also runs external agents that speak the Agent Client Protocol, and shift-alt-escape expands the message editor when you need room.
The docs don't mention voice or dictation. There is a microphone in Zed, but it belongs to collaboration. In channels, Zed shares your microphone with others in the call when you join, if macOS allows it, and the mute_on_join setting starts you muted. That's voice chat with people, not speech-to-text.
The feature request
Zed's issue "Dictation for Assistant (inline and chat-based)" was opened in August 2024. It asks for a microphone in the assistant and a way to see and edit the transcript before it goes to the model. The issue is marked Planned in Zed's community requests project, and several pull requests that tried to add speech-to-text have been closed. In other words, the Zed team has acknowledged it, but there's no shipped feature or date.
Apple Dictation in Zed
In the same thread, one commenter said macOS dictation works in the assistant once Dictation is turned on in System Settings. Another said they found it too error-prone and flaky. We'd treat it as worth a try in your version. Remember that it stops by itself after 30 seconds of silence, which can happen while you're reading a diff.
Why some tools can't find the agent input
An open issue from 3 October 2026 reports that one third-party dictation app transcribes fine but refuses to insert into Zed's agent input, saying no text field is focused. The app offers the transcript for manual pasting instead. The issue is labeled for macOS accessibility. A bot linked a similar report from another app built on Zed's GPUI framework. No cause has been confirmed yet.
A plausible explanation is that the tool looks for the focused field through the macOS accessibility interface before inserting, and doesn't find one in Zed's custom-drawn UI. A tool that pastes anyway, the way you would with ⌘V, avoids that particular check.
A setup that works now
- Open a thread in the Agent Panel and press
shift-alt-escapeif the prompt will be long. - Type your
@mentions first. File and symbol pickers are faster by keyboard than by voice. - Dictate the rest: the goal, what you saw, what must not change and how to check the result.
- Read it, fix misheard names and press Enter.
The order of a spoken prompt matters more than the tool. There's a longer guide in how to dictate a good prompt.
With FnFlow
FnFlow doesn't need to find the field before it pastes. You hold fn (or right ⌘), talk and let go. Whisper transcribes on your Mac, and FnFlow checks that you're still in the same app and window, then sends one ⌘V. Afterwards it tries to read the field back. Where an app doesn't expose the field, FnFlow says Paste sent instead of Pasted and keeps the text in history with its audio, so a failed paste is never silent.
We haven't run a dedicated test in Zed's Agent Panel, so this is what the design implies, not a measured result. If something doesn't land, copy it from history. For editors we have tested more, see voice dictation for Cursor and dictation in VS Code.
Questions
Does Zed have built-in dictation?
Not as of 18 October 2026. The Agent Panel docs don't mention voice, and the feature request "Dictation for Assistant" has been open since August 2024 with the status Planned. The microphone in Zed is for voice chat in collaboration channels.
Why does my dictation app say no text field is focused in Zed?
An open Zed issue from October 2026 reports exactly that for one dictation app in the Agent Panel input. The cause isn't confirmed. A likely reason is that the tool looks for the field through macOS accessibility and doesn't find it.
Does Apple Dictation work in Zed?
Reports in Zed's issue tracker are mixed. One commenter said it works in the assistant, another found it error-prone. Try it in your version before relying on it.
Will FnFlow work in Zed's Agent Panel?
FnFlow pastes with ⌘V, so it doesn't need to find the field first. We haven't run a dedicated test in Zed. If Zed doesn't expose the field, expect the pill to say Paste sent rather than Pasted, with the text kept in history.
Local Wispr Flow alternativeFnFlow vs VoiceInkVoice input for CursorHow 7 Mac dictation apps get you to your first dictationWhisper.cpp on Apple Silicon: which large-v3-turbo quantization to shipHow I built a local dictation appVoice dictation in Claude Code and the terminalWhy holding fn doesn't work for dictation on Mac (and how to fix it)How to run Whisper locally on an Apple Silicon MacSpokenly alternative for developersFnFlow vs EnviousWisprHow to dictate a good prompt for AI coding toolsVoice dictation in VS Code on a MacTeaching Whisper your jargon: initial prompt and replacementsWhy Whisper types "Thank you" on silence, and how to filter itPrivate dictation on a Mac at work: what actually leaves your machineVoice input for Codex CLI and the ChatGPT appVoice coding on a Mac: what actually worksWriting git commit messages by voiceDictating code review comments on GitHubDictation in IntelliJ IDEA, PyCharm and WebStorm on a MacDictation in Terminal, iTerm2 and WarpWhich Whisper model to use on an Apple Silicon MacPush-to-talk dictation on a Mac: hold, toggle or streamSuperwhisper alternative for developersClaude Code voice mode: how to enable /voice and fix itWhisper apps: five ways to run Whisper locallyTalon voice coding vs a dictation appPush to talk on a MacDoes Claude have a voice mode?VoiceInk vs Wispr FlowVoiceInk vs SuperwhisperParakeet vs Whisper on a MacApple SpeechAnalyzer vs WhisperWhisper.cpp vs faster-whisper on a MacMlx-whisper vs whisper.cppGitHub Copilot voice inputGemini CLI voice inputVoice input in Google AntigravityVibe coding with voice input