Standard Compute

All fixes · OpenAI Codex app · Quota & billing

Codex dictation usage limit reached: why and how to fix

Codex dictation usage limit reached / You're out of Codex and Work usage

Dictation in the Codex desktop app stops when the shared Codex and Work allowance runs out (OpenAI documents no separate dictation quota). Check Settings → Usage for the reset time, then wait, add credits or redeem a banked reset. To keep talking in the meantime, dictate in ChatGPT on the web or phone, or use your operating system's dictation (Dictation on macOS, Win+H on Windows).

What causes it

  • The Codex and Work allowance is spent: the desktop app shows “You're out of Codex and Work usage” and blocks features that run on it, which users report includes dictation.
  • Voice in the desktop app uses your existing Codex usage at $0.05 a minute, per OpenAI's Codex pricing page, so long voice sessions spend the same budget as coding.
  • Coding is usually what spent it: Codex CLI, the IDE extension, cloud tasks and the app all draw from one allowance, so a heavy coding day leaves nothing for dictation.
  • Regular Chat voice is separate: OpenAI says it has its own limits and does not consume Codex usage, which is why dictation on the web can work while the desktop app refuses.

How to fix it

  1. Open Settings → Usage in the app, or chatgpt.com/codex/settings/usage, to see which window is spent and when it resets.
  2. Wait for the reset, buy credits to continue now, or redeem a banked reset with /usage if your account has one.
  3. Dictate somewhere that does not use the Codex allowance: ChatGPT on the web or the mobile app, or your operating system's dictation into the prompt box.
  4. Move heavy coding off the allowance (an API key or another provider in Codex CLI) so the plan allowance lasts for the app features that need it.

Sources: OpenAI: Codex pricing (voice uses Codex usage); OpenAI Help: using Codex with your ChatGPT plan; OpenAI Help: dictation FAQ; openai/codex issue #46701; openai/codex issue #37619

Keep working on a fixed monthly plan

Standard Compute does not do dictation. What it can take over is the coding that usually spends the allowance: run Codex CLI on it with standardcompute launch codex, or a [model_providers] entry in ~/.codex/config.toml with base_url = "https://api.stdcmpt.com/v1", wire_api = "responses" and model standardcompute, and your ChatGPT allowance is left for the app, voice and dictation. Plans cost a fixed $19 to $249 a month for a set compute budget, with no 5-hour or weekly windows. Each key gets 300 to 2,000 requests a minute by plan, and at 100% of the budget new requests return HTTP 402 until renewal, with no overage charge.

Get your API key How it connects

Plans from $19 a month with $20 of compute. Eligible new accounts can try $0.25 of compute first, no card. Trial terms

FAQ

Does Codex have a separate dictation limit?

Not one OpenAI documents. Its dictation help covers how the feature works and how audio is stored; the limits it documents are for voice, which in the desktop app uses Codex usage. User reports tie the dictation block to the shared Codex and Work allowance.

Why does dictation work on chatgpt.com but not in the app?

Regular Chat voice and dictation on the web have their own limits and do not consume Codex usage, while the desktop app's Codex and Work features share one allowance. When that allowance is spent, the app blocks while the web keeps working.

Related errors