Voice transcription no-op

Resolved 💬 13 comments Opened Feb 25, 2026 by isaacjwilliams Closed Feb 25, 2026

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

What version of Codex CLI is running?

0.105.0

What subscription do you have?

Pro

Which model were you using?

_No response_

What platform is your computer?

Linux 6.18.12-200.fc43.x86_64 x86_64 unknown

What terminal emulator and version are you using (if applicable)?

kitty 0.45.0

What issue are you seeing?

Holding space to activate voice transcription simply types lots of spaces.

What steps can reproduce the bug?

Launch Codex and hold down the spacebar.

What is the expected behavior?

I speak and text is entered into the prompt box.

Additional information

I do have the voice transcription feature enabled in my ~/.codex/config.toml:

[features]
prevent_idle_sleep = true
voice_transcription = true

I am on Fedora 43 with KDE Plasma.

View original on GitHub ↗

13 Comments

github-actions[bot] contributor · 4 months ago

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

#12819
#12804

Powered by Codex Action

etraut-openai contributor · 4 months ago

Thanks for reporting. This feature isn't currently supported on Linux.

isaacjwilliams · 4 months ago

@etraut-openai Ah, my bad. I saw it in the release notes and was excited to try it, but did not realize that it was unsupported on Linux (I don't see that mentioned in the release notes). I do see the Under-development features enabled warning when I start Codex with the feature enabled, which contributed to my thought that it was supported on my platform.

d1egoaz · 4 months ago

~I have the same issue, I'm on Mac, using Ghostty @etraut-openai~

~Pressing and holding space just add more spaces.~

~I’m wondering if we can expose this option with a new command like /v, or let users choose their own shortcut and leave the space bar alone.~

It's working, you need to enable:

voice_transcription = true

mention your thread ID 019c9710-fa02-7433-8555-228fc50abc24 in an existing issue.

│  >_ OpenAI Codex (v0.105.0)                                           │
│                                                                       │
│ Visit https://chatgpt.com/codex/settings/usage for up-to-date         │
│ information on rate limits and credits                                │
│                                                                       │
│  Model:                gpt-5.3-codex (reasoning high, summaries auto) │
│  ...                                                                     │
│  Credits:              Unlimited                                      │
╰────────────────────────────────────────────

d1egoaz · 4 months ago

I have the same issue, I'm on Mac, using Ghostty @etraut-openai

nvm, I didn't have:

voice_transcription = true

It's working great!

d1egoaz · 4 months ago

I'd say the release notes are confusing:

You can now dictate prompts by holding the spacebar to record and transcribe voice input directly in the TUI. (https://github.com/openai/codex/pull/3381)

There is no comment about this needed to be enabled in the conf file.

d1egoaz · 4 months ago

For newcomers to Codex like me, this can be done via:

codex features enable voice_transcription

juanfradb · 4 months ago

+1 reproducible on Linux. Holding space inserts spaces instead of dictation, which blocks voice input in Codex CLI.

d1egoaz · 4 months ago

BTW
The releases notes were updated from:

You can now dictate prompts by holding the spacebar to record and transcribe voice input directly in the TUI. (https://github.com/openai/codex/pull/3381)

You can now dictate prompts by holding the spacebar to record and transcribe voice input directly in the TUI. This feature is still under development; to enable it set features.voice_transcription = true in your config. (https://github.com/openai/codex/pull/3381)

But here it's still outdated:
https://developers.openai.com/codex/changelog#github-release-290476287

umairshahid-1 · 4 months ago

I have encountered the same issue. I have enabled the feature in my config.toml file: voice_transcription = true. For context, I am on Windows 11, using the latest Codex version 0.105.0 with a pro subscription.

etraut-openai contributor · 4 months ago

Features that are marked "under development" are not yet ready for use. We'll mark it as "experimental" once it's ready for use. For details, refer to our documented feature maturity levels.

guidedways contributor · 4 months ago

This seems to work on macOS just that it stops working after a little while: https://github.com/openai/codex/issues/12891

atsyplenkov · 4 months ago

+1 reproducible on Linux. Just a bunch of spaces. I am on Ubuntu 24.04, codex v0.106.0 and tried various terminals including Ghostty and Konsole