Repository navigation
Requirements and Recommendations
▶ Watch in the Getting Started video: Before you install (2:21)
Read this before you install. It covers what you need to run the VideoExpress Scripting skill in Claude Desktop, ChatGPT and Grok, and what we recommend for the best results.
Short version: a paid AI plan is not strictly required to use the skill, but a full production (a long script, many prompts, and browser control) can use up a free plan's limits quickly, and the more capable models on paid plans write noticeably better scripts and prompts. We recommend a paid plan with access to the most capable models.
Plan features change often. The tables below reflect the Claude and ChatGPT pricing pages as of October 2026; check claude.com/pricing and chatgpt.com/pricing for the latest.
- VideoExpress (app.videoexpress.ai), a one-time purchase, for image and video generation. The skill writes the prompts; VideoExpress renders them.
- One AI assistant from the options below: Claude, ChatGPT or Grok.
- A desktop or laptop if you want to install from the plugin marketplace or have the assistant drive VideoExpress in the browser. Planning and prompt writing also work on web and mobile.
| What you want to do | Free | Paid (Pro, Max, Team, Enterprise) |
|---|---|---|
| Use the skill (upload the zip under Customize → Skills) | ✅ | ✅ |
| Install from the plugin marketplace in Claude Desktop (walkthrough) | ❌ plugins need a paid plan | ✅ |
| Let Claude drive VideoExpress in the browser (Claude in Chrome) | ❌ | ✅ |
| Access to the most capable models | ❌ | ✅ |
On Free: turn on Code execution and file creation under Settings → Capabilities, then upload videoexpress-scripting-claude.zip from the releases page under Customize → Skills. You can plan, write the script and get every prompt; you paste the prompts into VideoExpress yourself (the manual operating mode).
On a paid plan: install from the marketplace in Claude Desktop so the skill appears as a plugin, and add the Claude in Chrome extension if you want the assisted or automatic operating modes.
| What you want to do | Free | Go | Plus / Pro / Business |
|---|---|---|---|
| Install from the plugin marketplace in ChatGPT (walkthrough) | ✅ | ✅ | ✅ |
| Use the skill | ✅ (skills in beta) | ✅ | ✅ |
| Let ChatGPT drive VideoExpress in the browser (agent mode) | Limited | Limited | ✅ (expanded on Pro) |
| Advanced reasoning models | ❌ | ❌ | ✅ |
On Free or Go: you can install the plugin and use the skill, but usage limits are tight and you get the default model rather than the advanced reasoning models.
On Plus or above: expanded agent-mode usage and the advanced reasoning models. This is the recommended setup for a full production.
Grok skills are available on the free trial and paid plans only. Grok installs the skill itself from its GitHub folder link, so nothing is installed on your computer (walkthrough).
Letting Grok drive VideoExpress needs browser control. Where your Grok can't control a browser, use the manual operating mode: Grok writes the prompts and you paste them into VideoExpress.
A VideoExpress production is a long conversation. Intake, the Core Bible, a clip-by-clip script, a start-image prompt and a multishot prompt for every clip, then revisions as clips come back, adds up to a lot of text for the assistant to read and write. Browser control is heavier still: every click, screenshot and page read counts against your limit, and a single batch of takes involves dozens of them.
On a free plan, expect to hit the usage limit partway through a project, sometimes in the middle of a generation run, and to wait for it to reset before you can continue. Paid plans have several times the usage per session.
The skill relies on the model following a lot of detailed rules at once: pasting bible blocks word for word, describing the frame rather than the plan, tracking camera direction across chained clips, keeping sound transient, and working out which other clips a fix affects. The base or default models on free plans can follow the skill, but in practice they:
- write shorter, less specific scripts and prompts, with less of the detail VideoExpress needs to keep characters, outfits and locations consistent;
- are less aware of how the video environment reacts to a change, for example that a fix to one clip's ending changes the chain frame for the next, or that a new noun in a prompt is likely to appear on screen;
- are more likely to drop a rule in a long prompt, such as quoting dialogue twice or leaving a camera move without a destination, which costs you time on takes that fail.
The more capable models on paid plans hold more of the project in mind and catch these knock-on effects before you spend takes on them.
| Minimum that works | Recommended | |
|---|---|---|
| Claude | Free plan, skill uploaded as a zip, manual mode | Pro or above, installed from the marketplace in Claude Desktop, with Claude in Chrome |
| ChatGPT | Free plan, plugin installed from the marketplace, manual mode | Plus or above, plugin installed from the marketplace, agent mode for browser control |
| Model | Whatever your plan defaults to | The most capable model your plan offers |
| VideoExpress | A VideoExpress purchase | Time for development tests and re-rolls |
Next: install the skill.
See all styles