[Integrations](https://spiral.ai/integrations) › [Codex](https://spiral.ai/integrations/codex) › [LTX 2.3 Pro Retake](https://spiral.ai/integrations/codex/ltx-2-3-pro-retake)

# How to Use LTX 2.3 Pro Retake in Codex

LTX 2.3 Pro Retake revises a selected section of an existing video while preserving the rest. Unlike LTX 2.3 Fast, it supports targeted edits to footage and its soundtrack.

|  |  |
| --- | --- |
| Model | LTX 2.3 Pro Retake |
| Available through | fal |
| Category | Video |
| Price | $0.13 per second. Billed on `duration`, default 5. |
| Returns | video |
| Runs | In the background |
| Setup | About 2 minutes, no API key |

## Prompts to try with LTX 2.3 Pro Retake in Codex

- Use LTX 2.3 Pro Retake on assets/launch-demo.mp4 from second 2 through second 6. Replace picture and sound: a medium shot shows the presenter smiling, then saying, "Your next task starts here." Keep the camera static and the lighting soft. Save the result beside the original for review.
- Read the product color guidance in this repository, then use LTX 2.3 Pro Retake on seconds 0 through 4 of assets/product-shot.mp4. Request the approved product color while retaining the shot's composition and original soundtrack. Save a comparison clip.
- Use LTX 2.3 Pro Retake on seconds 3 through 7 of assets/mascot.mp4. Replace only the picture: the clay robot slowly raises one hand and waves once, with consistent soft lighting and a static camera. Save the result as a separate candidate.
- Use LTX 2.3 Pro Retake to revise seconds 1 through 5 of assets/teaser.mp4. Replace only the picture with a gentle camera push toward the product on the table. Keep the product centered and use one consistent window-light direction. Save the retake for visual review.
- Use LTX 2.3 Pro Retake on seconds 0 through 5 of assets/outro.mp4. Replace only the audio with a calm voice saying, "Pick a task. Make it happen." Request quiet room ambience underneath. Save the result and flag it for a listening check.

## How to set up LTX 2.3 Pro Retake 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 LTX 2.3 Pro Retake**: 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": "fal",
    "model": "fal-ai/ltx-2.3/retake-video",
    "input": {
      "video_url": "https://v3b.fal.media/files/b/0a90e00e/eNWXGtaGyZO--e6KcJh-3_171_video.mp4",
      "prompt": "Change flower to red rose"
    }
  }'
```

Then paste one of the prompts above: Codex finds LTX 2.3 Pro Retake on Spiral, shows the price and runs it.

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

## LTX 2.3 Pro Retake tools and prices

- [LTX Video 2.3 Pro · Retake Video](https://spiral.ai/tools/fal/fal-ai/ltx-2.3/retake-video) (fal): $0.13 per second. Billed on `duration`, default 5.

## Settings you can ask for

| Setting | Allowed values | Default | Description |
| --- | --- | --- | --- |
| start_time | 0 to 20 | 0 | The start time of the video to retake in seconds |
| duration | 2 to 20 | 5 | The duration of the video to retake in seconds |
| retake_mode | replace_audio, replace_video, replace_audio_and_video | replace_audio_and_video | The retake mode to use for the retake |

## What can you do with LTX 2.3 Pro Retake in Codex?

Revise short sections of demo and launch videos using the same project notes that guide development. A coding agent can turn those notes into targeted retake requests and save comparison clips for review.

### Product color tests

Request a different product color in a short section of an existing showcase clip, then compare the result with the original composition.

### Performance revisions

Try a different visible reaction in an existing scene, such as a slow smile or a hesitant glance. Describe the gesture rather than an abstract emotion.

### Short dialogue revisions

Request a new spoken line in a selected section of an existing clip. Regenerate picture and sound together when the performance also needs to change.

### Soundtrack alternatives

Try a different sound treatment for an existing scene while retaining its picture. Audition the result carefully, especially when generating ambience without speech.

### Camera movement studies

Request a gentle push-in or a static-camera alternative for a short section of existing footage. Compare how each treatment directs attention to the subject.

### Lighting alternatives

Test a different lighting treatment within an existing shot, such as soft window light or a single warm lamp. Keep the lighting instructions consistent.

### Animation revisions

Request a simpler action in a selected section of an existing animated clip. Keep the subject and visual style explicit so the result is easy to compare.

## Overview

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

## LTX 2.3 Pro Retake in other agents

- [Amp](https://spiral.ai/integrations/amp/ltx-2-3-pro-retake)
- [ChatGPT](https://spiral.ai/integrations/chatgpt/ltx-2-3-pro-retake)
- [Claude](https://spiral.ai/integrations/claude-desktop/ltx-2-3-pro-retake)
- [Claude Code](https://spiral.ai/integrations/claude-code/ltx-2-3-pro-retake)
- [Cline](https://spiral.ai/integrations/cline/ltx-2-3-pro-retake)
- [Cursor](https://spiral.ai/integrations/cursor/ltx-2-3-pro-retake)
- [Devin](https://spiral.ai/integrations/devin/ltx-2-3-pro-retake)
- [dots](https://spiral.ai/integrations/dots/ltx-2-3-pro-retake)
- [Factory Droid](https://spiral.ai/integrations/factory-droid/ltx-2-3-pro-retake)
- [Gemini](https://spiral.ai/integrations/gemini/ltx-2-3-pro-retake)
- [Gemini CLI](https://spiral.ai/integrations/gemini-cli/ltx-2-3-pro-retake)
- [Google Antigravity](https://spiral.ai/integrations/antigravity/ltx-2-3-pro-retake)
- [goose](https://spiral.ai/integrations/goose/ltx-2-3-pro-retake)
- [Grok](https://spiral.ai/integrations/grok/ltx-2-3-pro-retake)
- [Grok Build](https://spiral.ai/integrations/grok-build/ltx-2-3-pro-retake)
- [Hermes](https://spiral.ai/integrations/hermes/ltx-2-3-pro-retake)
- [Instinct](https://spiral.ai/integrations/instinct/ltx-2-3-pro-retake)
- [Manus](https://spiral.ai/integrations/manus/ltx-2-3-pro-retake)
- [Muse](https://spiral.ai/integrations/muse/ltx-2-3-pro-retake)
- [Muse Code](https://spiral.ai/integrations/muse-code/ltx-2-3-pro-retake)
- [n8n](https://spiral.ai/integrations/n8n/ltx-2-3-pro-retake)
- [Notion](https://spiral.ai/integrations/notion/ltx-2-3-pro-retake)
- [OpenClaw](https://spiral.ai/integrations/openclaw/ltx-2-3-pro-retake)
- [opencode](https://spiral.ai/integrations/opencode/ltx-2-3-pro-retake)
- [Perplexity](https://spiral.ai/integrations/perplexity/ltx-2-3-pro-retake)
- [Raycast](https://spiral.ai/integrations/raycast/ltx-2-3-pro-retake)
- [VS Code](https://spiral.ai/integrations/vscode/ltx-2-3-pro-retake)
- [Windsurf](https://spiral.ai/integrations/windsurf/ltx-2-3-pro-retake)
- [Zapier](https://spiral.ai/integrations/zapier/ltx-2-3-pro-retake)

## More Codex guides

- [LTX 2.3 Pro Image to Video](https://spiral.ai/integrations/codex/ltx-2-3-pro-image-to-video)
- [LTX 2.3 Pro Text to Video](https://spiral.ai/integrations/codex/ltx-2-3-pro-text-to-video)
- [LTX 2.3 Pro Audio to Video](https://spiral.ai/integrations/codex/ltx-2-3-pro-audio-to-video)
- [LTX 2.3 Reframe](https://spiral.ai/integrations/codex/ltx-2-3-reframe)
- [LTX 2.3 Fast Text to Video](https://spiral.ai/integrations/codex/ltx-2-3-fast-text-to-video)
- [LTX 2.5 Alpha Gen](https://spiral.ai/integrations/codex/ltx-2-5-alpha-gen)

## Questions about LTX 2.3 Pro Retake in Codex

**Can Codex use LTX 2.3 Pro Retake?**

Yes. Add Spiral to Codex as an MCP server, and Codex can find and run LTX 2.3 Pro Retake like any built-in tool.

**How much does LTX 2.3 Pro Retake cost in Codex?**

$0.13 per second. Billed on `duration`, default 5. 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 a fal account to use LTX 2.3 Pro Retake in Codex?**

No. Spiral runs LTX 2.3 Pro Retake on its own fal account. You sign in to Spiral once and never paste a key.

**How long does it take to set up LTX 2.3 Pro Retake in Codex?**

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

**Does Codex wait while LTX 2.3 Pro Retake runs?**

LTX 2.3 Pro Retake runs in the background. Codex starts the run, checks on it and gives you the video when it's ready.

**What is LTX 2.3 Retake useful for?**

Use it when a short section needs another attempt: a different visual treatment, a revised performance, or new sound. It starts from an existing video and regenerates the selected section.

**Can I change the picture while keeping the original soundtrack?**

Yes. Replace only the video when the existing sound already works; replace both when the picture and sound need to change together. Retake preserves the content outside the selected section.

**What resolution does LTX 2.3 Retake produce?**

The documented output is 1920×1080 landscape or 1080×1920 portrait. Input footage is center-cropped to that aspect ratio, so check subjects near the edges.

**What should I put in a retake prompt?**

Describe the desired shot, the subject's action, the lighting, and any camera movement. Keep the scene focused, describe sound when replacing it, and put any spoken line in quotation marks.

**What should I check before accepting a retake?**

Check whether the requested details actually appeared, then watch the edited section in context. Listen closely to sound-only passages: the model card warns that generated audio without speech may have lower quality.

