Feature request: Add native microphone and voice dictation to the Codex chat input

Resolved 💬 5 comments Opened Jul 16, 2026 by Yojimbos Closed Aug 6, 2026

💡 Likely answer: A maintainer (github-actions[bot], contributor) responded on this thread — see the highlighted reply below.

Summary

Please add an optional microphone / voice dictation control directly to the Codex chat input, especially in the IDE extension and follow-up workflow used after reviewing generated changes.

Problem

When working with Codex on real projects, follow-up instructions are often detailed and iterative: describe a failed test, clarify expected behavior, list several files to update, or explain a CI/CD scenario. Typing these instructions repeatedly slows down the feedback loop, particularly when the user already has the relevant code and diff open in the editor.

Today, users must rely on external speech-to-text tools, dictate into another application, and then copy the text into Codex. This breaks focus, adds friction, and often provides poor support for languages other than English.

Proposed experience

Add a microphone button near the Codex message composer and the follow-up changes input.

Possible behavior:

Click or hold to record, then convert speech to editable text before submission.
Support push-to-talk and an optional keyboard shortcut.
Let the user review and edit the transcription before sending it.
Support multilingual dictation and automatic language detection, including Ukrainian.
Clearly indicate when recording is active.
Keep the feature optional and disabled until the user activates it.
Provide transparent privacy controls, including whether audio is processed locally or remotely and whether recordings are retained.

Why this would help

Faster iteration when reviewing and requesting follow-up changes.
Less context switching between the IDE, dictation software, and clipboard.
Better accessibility for users who cannot comfortably type long prompts.
More natural interaction for complex engineering explanations, debugging context, and CI/CD investigations.
Stronger usability differentiation for Codex compared with coding assistants that only support keyboard input.
Potentially higher product engagement because users can provide richer context with less effort.

Example workflow

Codex proposes changes and shows the diff.
The user presses the microphone button in the follow-up field.
The user says: “The implementation is mostly correct, but keep the existing pipeline parameter names, update only the Enlight validation stage, and add a fallback when the previous major version is unavailable.”
Codex transcribes the request into editable text.
The user corrects any technical names and submits it.

This would make iterative work with Codex considerably faster without changing the underlying agent workflow.

View original on GitHub ↗

5 Comments

github-actions[bot] contributor · 1 month ago

Potential duplicates detected. Please review them and close your issue if it is a duplicate.

#33221

Powered by Codex Action

Yojimbos · 1 month ago

Environment

VS Code: 1.122.1
OS: Windows 11 Enterprise 25H2, build 26200.8457, AMD64
Codex IDE extension: openai.chatgpt 26.707.71524
Codex package: 0.144.2

Acceptance criteria

Add an optional microphone button to:
the main Codex prompt input;
the follow-up Request changes input shown after code generation, diff review, or implementation review.

Provide both:
push-to-talk behaviour;
a configurable keyboard shortcut, unassigned by default.

While recording, display:
a clear recording indicator;
elapsed recording time;
one-click Stop and Cancel actions.

Insert the transcription into the currently active Codex input field as editable text.

Never automatically submit the transcribed prompt.

Allow the user to continue typing before, during, or after transcription without blocking or replacing manually entered text.

Support automatic language detection and an optional language override, for example:
Auto
English
Ukrainian
other supported languages

Support Ukrainian transcription end-to-end, including punctuation and Cyrillic text.

Preserve common technical terms, filenames, paths, package names, commands, and identifiers whenever possible.

Provide clear microphone permission and privacy controls:
first-use permission prompt;
setting to disable voice input entirely;
clear information about whether audio is processed locally or remotely;
clear information about whether audio recordings are retained.

Handle common failures gracefully:
no microphone available;
microphone permission denied;
network or transcription failure;
interrupted recording;
partial transcription;
retry without losing already transcribed text.

Cancelling a recording must not remove text that was already present in the prompt field.

Voice input must not interfere with normal keyboard input, editor shortcuts, or Codex actions.

Nice-to-have enhancements

Automatic punctuation toggle.
Visual microphone level meter.
Configurable press-and-hold versus toggle recording mode.
Custom vocabulary or recognition hints for technical terminology.
A shortcut to start dictation directly in the currently focused Codex input.
Optional local transcription mode when supported by the platform.

Yojimbos · 1 month ago

Thanks for flagging #33221. I reviewed it, and this request is related but not a duplicate.

#33221 asks for a full conversational voice mode in the Codex Desktop app, mainly for spoken interview preparation and project walkthroughs.

This issue is specifically about speech-to-text dictation inside the Codex IDE extension and its existing text inputs, including the follow-up Request changes field. The expected result is editable transcribed text that is never auto-submitted, with multilingual support, keyboard/push-to-talk controls, and privacy handling.

The two requests may share underlying audio/transcription infrastructure, but they target different products, interaction models, and acceptance criteria. I suggest keeping this issue open and linking the implementation where shared components are relevant.

Yojimbos · 22 days ago

@etraut-openai, thank you for reviewing the request.

Could you please clarify whether the realtime voice mode and microphone controls shipped for Codex Desktop are also planned for the Codex IDE extension for VS Code?

The requirement in this issue is narrower and differs from the Desktop conversational voice mode described in #33221.

We are specifically requesting speech-to-text dictation inside the existing Codex IDE extension inputs:

the main prompt field;
the follow-up Request changes field after reviewing generated code or a diff;
transcription inserted as editable text;
no automatic submission;
multilingual speech recognition, including Ukrainian;
optional push-to-talk and keyboard controls.

The Desktop implementation may provide reusable voice infrastructure, but it does not appear to confirm support for these IDE-extension workflows.

Could you please confirm one of the following?

This functionality is already available in the VS Code extension, and where it can be enabled.
It is planned as part of the Desktop voice implementation rollout to the IDE extension.
It is tracked in another canonical issue.
If none of the above applies, could this issue please be reopened as an IDE-extension-specific feature request?

We would mainly like to understand whether the original IDE requirement is covered and where its future progress can be tracked.

Yojimbos · 7 days ago

@etraut-openai, following up on the IDE-specific dictation request now that we have additional clarification from OpenAI Support.

OpenAI Support case #12971658 confirmed that this request has been logged and escalated as product feedback for the Codex IDE/VS Code extension area. Support also confirmed that there is currently no public documentation establishing that editable dictation should already be available in the VS Code extension, and that the completed Desktop conversational Voice feature does not by itself establish support for this IDE workflow.

I also rechecked the openai/codex issue tracker today and still could not find another canonical issue clearly tracking editable speech-to-text dictation for the VS Code extension.

The requested behavior remains specifically:

speech-to-text in the existing Codex IDE input fields;
editable transcription before sending;
no automatic submission;
support in both the main prompt and Request changes fields;
multilingual recognition, including Ukrainian;
optional microphone / push-to-talk / keyboard controls.

Given that #33221 tracked Desktop conversational voice, while this IDE-specific requirement remains separately logged with Support and does not appear to have another public tracker, could you please reconsider the duplicate classification for #33548?

If there is another canonical issue for this VS Code requirement, please link it here. Otherwise, reopening #33548 would provide a clear public tracking point for the IDE-specific feature request.

Thank you.