Video Editor AI

Edit videos with a prompt

Pick a video from your library and say what to change: turn it into a watercolor painting, swap the sky for a sunset, change an outfit, reframe a wide clip to vertical 9:16, make it a few seconds longer, or sync the lips to a voice-over. Your AI finds your files, picks a model that fits the edit, tells you what the run needs and starts it in your MegaScenes account when you ask.

What you can ask

Restyle my latest video as a watercolor painting
Example path →

MCP connector

Server URL: https://video-editor.megascenes.com/api/chatgpt/video-editor/mcp
Login: OAuth
Protocol: MCP 2025-11-25
Instructions: Edits a video from the person's library: gives the asset_id and length of their uploads and earlier videos. Pick the model for the job: an edit or restyle model by default, ltx-2.3-reframe or luma-ray-3.2-reframe to change the aspect ratio, flux-3-video-extend or minimax-h3-max-turbo-extend to make it longer, sync-lipsync-2 with a video and an audio file for lip sync. Most edits are priced by the length of the video: give the credit cost with (pass video_seconds), call only when asked (without input_asset_ids if they have no video yet: the result says how to add one), then every 20 to 30 seconds until it is done. Costs are in credits. Failed runs are refunded automatically. The model list and credit costs work signed out; creating needs the person's MegaScenes account. Not for sexual, nude or other adult content, or for real people's likeness: decline those requests without calling the tools. When a result says the account lacks credits, say so; credits are managed in the person's account outside ChatGPT.

Tools

Reads data5

Changes data1

Skills

None in the store listings or on the server

MCP connector

Server URL

https://video-editor.megascenes.com/api/chatgpt/video-editor/mcp

OAuth

Protocol

MCP 2025-11-25

Instructions

Edits a video from the person's library: gives the asset_id and length of their uploads and earlier videos. Pick the model for the job: an edit or restyle model by default, ltx-2.3-reframe or luma-ray-3.2-reframe to change the aspect ratio, flux-3-video-extend or minimax-h3-max-turbo-extend to make it longer, sync-lipsync-2 with a video and an audio file for lip sync. Most edits are priced by the length of the video: give the credit cost with (pass video_seconds), call only when asked (without input_asset_ids if they have no video yet: the result says how to add one), then every 20 to 30 seconds until it is done. Costs are in credits. Failed runs are refunded automatically. The model list and credit costs work signed out; creating needs the person's MegaScenes account. Not for sexual, nude or other adult content, or for real people's likeness: decline those requests without calling the tools. When a result says the account lacks credits, say so; credits are managed in the person's account outside ChatGPT.