Product

Edit videos with ChatGPT

Connect Vyra to ChatGPT over MCP. ChatGPT reads your transcript, cuts the timeline, adds captions, and exports, all from the chat.

By Sulan Zhang, co-founder of Vyra · Updated 2026-10-29

Real examples from @sulansart

How it works

Add the Vyra MCP server. On the web, go to https://chatgpt.com/plugins, click the plus button, then "Add custom MCP server". Name it Vyra, enter https://api.usevyra.com/mcp under Connection, set authentication to OAuth, and select "Create as a plugin". Install it and sign in to Vyra when the OAuth window opens. For Codex, including in the ChatGPT desktop app, run these instead.
```
codex mcp add vyra --url https://api.usevyra.com/mcp
codex mcp login vyra
```
Ask the agent to open a project or create one. Nothing has to stay open on your side. The agent edits on Vyra's servers, and the project shows up in the web and iPhone apps whenever you want to look.
Upload your footage. Each clip gets transcribed and described.
Tell ChatGPT what to make. Length, platform, keep, cut, captions, music.
Watch the timeline update. Ask for a read-back, then fix one thing at a time.

What you can ask for

You say	ChatGPT does in Vyra
"What's in this project?"	Lists clips with scene summaries
"Write a 60-second script from the best lines, then cut to it"	Scripts from the transcript, then assembles
"Caption everything word by word, white, above center"	Generates styled captions
"Put the screen recording top right while I talk about the prompt"	Places a picture-in-picture
"Cut to the beat of the attached track"	Lands cuts on the beats
"Export a 9:16 and a 1:1 version"	Renders both

Prompts to paste

Connect to my Vyra project "product demo". Summarize the footage, then propose a 30-second TikTok: hook in the first 2 seconds, one clean demo, one result shot. Wait for my OK before you edit.

Build the cut as planned. Remove pauses over half a second, filler words, and retakes. Add word-by-word captions, bold, white, slightly above center. No music. Read the timeline back with start and end times.

This project has ten takes of the same talking head. Keep the best take of each point, in the order I filmed them, target 45 seconds. Add 3-5 word captions and a lo-fi track at 25%. Export 1080x1920.

Compared to the usual way

	Traditional editor	ChatGPT with a passthrough integration	ChatGPT with Vyra
ChatGPT's role	None	Forwards your prompt to the vendor's AI	Plans and executes with real editing tools
Sees the footage	No	No	Transcripts, scene descriptions, frames
Where you work	The app	The app plus ChatGPT	ChatGPT, nothing else open
Cost	App subscription	App plus ChatGPT	Vyra MCP plan plus the ChatGPT plan you already have

Limits

ChatGPT edits what you uploaded. Vyra doesn't generate footage.
ChatGPT sometimes says a step is done early. Ask it to read the timeline back.

Pricing

Plan	Price	AI	Export
MCP Starter	$24/mo	Bring your own Claude, ChatGPT, or Cursor subscription	Up to 1440p
MCP Pro	$65/mo	Bring your own subscription	Up to 4K
Vyra AI Starter	$54/mo	Built-in AI, no external subscription needed	Up to 1440p
Vyra AI Pro	$129/mo	Built-in AI	Up to 4K

3-day free trial, then plans from $24/mo. See /pricing for the full breakdown.

FAQ

Can I use my existing ChatGPT account? Yes. That's what the MCP plans are for.

Is this the same as CapCut's Codex integration? No. That one only makes a draft inside CapCut. Vyra gives ChatGPT the full editing toolset, and it works with Claude and Cursor too.

Can I switch between ChatGPT and Vyra's built-in agent? Yes. Same project, same timeline.

/mcp
/docs/mcp
/prompts/how-to-prompt-vyra
/compare/vyra-vs-descript