[Integrations](https://spiral.ai/integrations) › [Codex](https://spiral.ai/integrations/codex) › [GPT-4o mini Transcribe](https://spiral.ai/integrations/codex/gpt-4o-mini-transcribe)

# How to Use GPT-4o mini Transcribe in Codex

GPT-4o mini Transcribe turns recorded speech into text for voice memos, interviews, and call reviews. OpenAI reports better language recognition and accuracy than the original Whisper models.

|  |  |
| --- | --- |
| Model | GPT-4o mini Transcribe |
| Available through | OpenAI |
| Category | Audio |
| Price | Input audio and prompt: $1.625 per 1M tokens. Output text: $6.50 per 1M tokens. |
| Returns | json |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with GPT-4o mini Transcribe in Codex

- Use GPT-4o mini Transcribe on recordings/bug-report.wav, using product names from README.md as spelling context. Then draft reproduction steps from the transcript.
- Transcribe recordings/walkthrough.mp3 with GPT-4o mini Transcribe. Supply component names from this project as context, then draft a documentation outline.
- Use GPT-4o mini Transcribe on recordings/design-review.wav. Split long audio between sentences, then compare the transcript's proposed changes with the current implementation.
- Use GPT-4o mini Transcribe to transcribe recordings/release-notes.m4a with package names from package.json as context. Then draft release notes without adding unmentioned changes.
- Transcribe recordings/test-feedback.wav with GPT-4o mini Transcribe. Then create a checklist of reported failures, preserving the transcript beside each item.

## How to set up GPT-4o mini Transcribe in Codex

Pick the way that suits you. Each takes about two minutes.

### Use Spiral's MCP server

1. **Add the hosted server**

```bash
codex mcp add spiral --url https://api.spiral.ai/mcp
```

2. **Sign in**: Codex opens the browser: sign in to Spiral and click Authorize. No API key. `codex mcp list` then shows Spiral.

```bash
codex mcp login spiral
```

### Use the Spiral skill

1. **Send Codex this message**: Your agent reads Spiral's skill and sets itself up.

```
set up https://spiral.ai/SKILL.md
```

2. **Sign in or paste a key**: Approve Spiral in the browser, or give the agent a key from the console.

### Use the Spiral CLI

1. **Install and sign in**: Installs the Spiral CLI, signs you in and adds its skill to your coding agents.

```bash
npx -y spiral-cli@latest init
```

### Call the Spiral API

1. **Create an API key**: In the console, under API keys.

2. **Run GPT-4o mini Transcribe**: One request; the result comes back by polling or a webhook.

```bash
curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "openai",
    "model": "gpt-4o-mini-transcribe",
    "input": {
      "file": "<file>"
    }
  }'
```

Then paste one of the prompts above: Codex finds GPT-4o mini Transcribe on Spiral, shows the price and runs it.

Setup guide in the docs: https://spiral.ai/docs/quickstarts/codex

## GPT-4o mini Transcribe tools and prices

- [GPT-4o mini Transcribe](https://spiral.ai/tools/openai/gpt-4o-mini-transcribe) (OpenAI): Input audio and prompt: $1.625 per 1M tokens. Output text: $6.50 per 1M tokens.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| prompt |  | — | An optional text to guide the model's style or continue a previous audio segment. The prompt should match the audio language. |
| include |  | — | Additional information to include in the transcription response. `logprobs` will return the log probabilities of the tokens in the response to understand the model's confidence in the transcription. |
| language |  | — | The language of the input audio. Supplying the input language in ISO-639-1 (e.g. `en`) format will improve accuracy and latency. |
| temperature | 0 to 1 | 0 | The sampling temperature, between 0 and 1. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. |

## What can you do with GPT-4o mini Transcribe in Codex?

Turn spoken bug reports and walkthroughs into text your coding agent can compare with the project. Use repository terminology as transcription context, then have the agent draft issues or documentation from the transcript.

### Voice memos into task drafts

Transcribe spoken reminders, then have your agent extract tasks while keeping the original wording available for review.

### Customer call review

Turn recorded calls into transcripts your agent can search for reported problems, requested changes, and follow-up commitments.

### Interview research

Create interview transcripts for finding quotes and recurring themes. Supply participant and organization names as spelling context.

### Podcast editorial preparation

Transcribe podcast sections, then have your agent draft episode notes from the text. Review selected quotes against the recording.

### Meeting follow-up

Create a transcript for drafting decisions and action items. Confirm ownership separately because the output does not identify speakers.

### Spoken bug reports

Transcribe recorded bug descriptions with project terminology supplied as context, then have your agent draft reproduction steps.

## Overview

- [OpenAI](https://spiral.ai/providers/openai)
- [All Codex guides](https://spiral.ai/integrations/codex)

## GPT-4o mini Transcribe in other agents

- [Amp](https://spiral.ai/integrations/amp/gpt-4o-mini-transcribe)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/gpt-4o-mini-transcribe)
- [Claude](https://spiral.ai/integrations/claude-desktop/gpt-4o-mini-transcribe)
- [Claude Code](https://spiral.ai/integrations/claude-code/gpt-4o-mini-transcribe)
- [Cline](https://spiral.ai/integrations/cline/gpt-4o-mini-transcribe)
- [Cursor](https://spiral.ai/integrations/cursor/gpt-4o-mini-transcribe)
- [Devin](https://spiral.ai/integrations/devin/gpt-4o-mini-transcribe)
- [dots](https://spiral.ai/integrations/dots/gpt-4o-mini-transcribe)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/gpt-4o-mini-transcribe)
- [Gemini](https://spiral.ai/integrations/gemini/gpt-4o-mini-transcribe)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/gpt-4o-mini-transcribe)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/gpt-4o-mini-transcribe)
- [goose](https://spiral.ai/integrations/goose/gpt-4o-mini-transcribe)
- [Grok](https://spiral.ai/integrations/grok/gpt-4o-mini-transcribe)
- [Grok Build](https://spiral.ai/integrations/grok-build/gpt-4o-mini-transcribe)
- [Hermes](https://spiral.ai/integrations/hermes/gpt-4o-mini-transcribe)
- [Instinct](https://spiral.ai/integrations/instinct/gpt-4o-mini-transcribe)
- [Manus](https://spiral.ai/integrations/manus/gpt-4o-mini-transcribe)
- [Muse](https://spiral.ai/integrations/muse/gpt-4o-mini-transcribe)
- [Muse Code](https://spiral.ai/integrations/muse-code/gpt-4o-mini-transcribe)
- [n8n](https://spiral.ai/integrations/n8n/gpt-4o-mini-transcribe)
- [Notion](https://spiral.ai/integrations/notion/gpt-4o-mini-transcribe)
- [OpenClaw](https://spiral.ai/integrations/openclaw/gpt-4o-mini-transcribe)
- [opencode](https://spiral.ai/integrations/opencode/gpt-4o-mini-transcribe)
- [Perplexity](https://spiral.ai/integrations/perplexity/gpt-4o-mini-transcribe)
- [Raycast](https://spiral.ai/integrations/raycast/gpt-4o-mini-transcribe)
- [VS Code](https://spiral.ai/integrations/vscode/gpt-4o-mini-transcribe)
- [Windsurf](https://spiral.ai/integrations/windsurf/gpt-4o-mini-transcribe)
- [Zapier](https://spiral.ai/integrations/zapier/gpt-4o-mini-transcribe)

## More Codex guides

- [GPT-4o mini TTS](https://spiral.ai/integrations/codex/gpt-4o-mini-tts)
- [Gemini 2.5 Flash TTS](https://spiral.ai/integrations/codex/gemini-2-5-flash-tts)
- [F5-TTS](https://spiral.ai/integrations/codex/f5-tts)
- [Gemini 2.5 Pro Preview TTS](https://spiral.ai/integrations/codex/gemini-2-5-pro-preview-tts)
- [Elevenlabs Tts Eleven V3](https://spiral.ai/integrations/codex/elevenlabs-tts-eleven-v3)
- [Gemini 3.1 Flash TTS](https://spiral.ai/integrations/codex/gemini-3-1-flash-tts)

## Questions about GPT-4o mini Transcribe in Codex

**Can Codex use GPT-4o mini Transcribe?**

Yes. Add Spiral to Codex as an MCP server, and Codex can find and run GPT-4o mini Transcribe like any built-in tool.

**How much does GPT-4o mini Transcribe cost in Codex?**

Input audio and prompt: $1.625 per 1M tokens. Output text: $6.50 per 1M tokens. You pay from your Spiral balance as you go, with no subscription, and a failed run is not charged. Your Codex plan is separate.

**Do I need an OpenAI account to use GPT-4o mini Transcribe in Codex?**

No. Spiral runs GPT-4o mini Transcribe on its own OpenAI account. You sign in to Spiral once and never paste a key.

**How long does it take to set up GPT-4o mini Transcribe in Codex?**

About two minutes: add the hosted server, then sign in.

**Does Codex wait while GPT-4o mini Transcribe runs?**

GPT-4o mini Transcribe runs in the background. Codex starts the run, checks on it and gives you the json when it's ready.

**What is GPT-4o mini Transcribe useful for?**

Use it to turn recorded speech into text for interviews, voice memos, podcasts, and customer call reviews. Your agent can then use the transcript to draft notes or extract follow-up tasks.

**How does GPT-4o mini Transcribe compare with Whisper?**

OpenAI reports better language recognition and transcription accuracy than the original Whisper models. It also reports improved reliability with accents, noisy recordings, and varying speaking speeds.

**How do I improve technical terms and name spellings?**

Provide the expected names, acronyms, and subject vocabulary as transcription context. Write that context in the recording's language and use the spellings you want preserved.

**Can GPT-4o mini Transcribe handle long recordings?**

Its maximum output is 2,000 tokens, so plan to transcribe long recordings in shorter sections. Split between sentences to preserve context.

**Does GPT-4o mini Transcribe produce subtitles or identify speakers?**

It returns a JSON transcript without native subtitle files or speaker annotations. It also lacks native word timestamps, so timed captions require a separate alignment step.

