How to Edit Videos with ChatGPT: A Step-by-Step Guide

Turning a folder of recordings into a finished video involves plenty of repetitive work. You review takes, remove false starts, assemble the sequence, find supporting shots, time the titles, and make another round of changes after watching it back.

Working with ChatGPT can take some of that grunt work off your plate, but it needs to be connected to a video editor through MCP. It can then use the editor’s tools to act on your instructions and shape the edit.

You can use any editor that offers a compatible MCP server, though the setup and available tools may differ. If you are looking for a free alternative, invideo Editor offers an MCP connection that lets ChatGPT perform the editing tasks you would otherwise do manually on a professional-grade timeline. It can also generate media for the project. However, generating AI images, video, or audio uses invideo credits separately.

This guide walks through building your first edit using ChatGPT on invideo Editor, from connecting your editor and assembling your first cut to reviewing and exporting the finished video.

We’ll use a two-minute product demonstration as an example. The prompts illustrate how to build an edit in stages, with a review between the major decisions.

1. Connecting ChatGPT to invideo Editor

Open invideo’s connection page, choose ChatGPT, and follow the setup instructions for your client. Complete the connection and sign in to your invideo account when prompted.

If you’re setting up the connection manually, invideo’s MCP server URL is:

https://mcp.invideo.io/mcp

ChatGPT supports connecting custom MCP servers through its plugin setup. Availability in a managed workspace depends on the permissions enabled for your account. OpenAI’s connection guide explains that setup.

Once connected, you can ask ChatGPT to work on an existing invideo Editor project or start a new one.

2. Set up your project and explain the video you have in mind

Upload your recordings to invideo Editor, or ask ChatGPT to upload the footage to your invideo project through the connection. For an existing project, share its link or identify it by name. If it contains several timelines, specify the one you want to work on.

Next, explain the concept: who the video is for, what it should show, and how you want it to feel.

For example:

We’re building a two-minute walkthrough of this coffee maker for people using it for the first time. Start with a brief shot of the finished coffee, then explain setup, brewing, and cleaning. Use the presenter’s explanation as the main thread, with close-ups wherever viewers need to see a specific action. Keep the pace clear and unhurried.

That gives ChatGPT the context for the decisions that follow. You can then give detailed instructions for the first stage, without asking it to finish every part of the video at once.

3. Ask ChatGPT to build the rough cut

Start with the structure and the core explanation. Tell ChatGPT how to handle repeated takes, pauses, and material that does not belong in the final video.

Build the base cut from these recordings. Follow the setup, brewing, and cleaning sequence. Remove false starts, repeated explanations, and long pauses, while keeping complete instructions and natural breaths. Use the takes where the presenter completes the explanation clearly. Leave music, titles, and effects for later.

ChatGPT can review the material and assemble the sequence on your timeline. You don’t need to mark every unwanted pause or retake yourself.

Watch the resulting cut before adding finishing elements. Check whether the opening establishes the purpose, the instructions follow a useful order, and each demonstration stays on screen long enough to understand.

For a closer look at this stage, see our guide to building a rough cut with AI.

4. Refine the sequence with specific feedback

Once the base cut is in place, direct the changes you want. Identify the relevant moment using a timestamp, a clip description, or a section name.

For example:

In the setup section, keep the close-up of the water tank on screen until the presenter finishes explaining the fill line. Shorten the repeated explanation before brewing starts. Leave the cleaning section unchanged.

This gives ChatGPT both the change and its boundary. You can make several related requests together, provided it is clear which part of the edit each one concerns.

You can also adjust a cut manually in invideo Editor. ChatGPT rereads the timeline as it works, so subsequent instructions use the updated edit. Creative context that isn’t visible on the timeline, such as why a particular take was approved, is still worth explaining in the conversation.

See a quick demonstration of a creator building and refining a travel edit with ChatGPT and invideo Editor.

5. Add the supporting shots your edit needs

As you review the sequence, you may find sections that would benefit from a closer view or a different illustration.

Ask ChatGPT to find the specific material and place it where it belongs:

Find the clearest close-up of the filter being inserted. Place it over the presenter’s explanation of that step, keeping the dialogue underneath.

The assistant can search footage in the project’s notebook or timeline and bring the relevant selection into the edit. This is useful when you know what a section needs but don’t want to search through every recording to find it.

If the project needs material you haven’t filmed, you can ask it to source stock footage or generate supporting visuals. Be precise about the purpose of the shot and whether it should come from your recordings, stock, or generation. For a product demonstration, use the actual product footage wherever viewers need to see how it works.

Generating new images, video, or audio uses invideo credits.

6. Build the sound around the edit

With the main sequence established, tell ChatGPT what the audio should do. That might involve finding background music, cleaning up the dialogue, or adding effects at particular moments.

Find a warm, understated background track and add it to the timeline. Keep the presenter easy to hear throughout. Let the music come forward over the opening coffee shot, then lower it when the explanation starts. Fade it out under the closing shot.

You can give further instructions for individual moments:

Keep the sound of the coffee pouring audible. Reduce the background noise in the cleaning section without making the voice sound thin.

Listen back to the transitions between clips as well as the individual sections. If something needs changing, identify the moment and describe the result you want. You can also adjust track levels yourself using invideo Editor’s audio mixer.

See the audio mixer and level controls in Devesh Sharma’s invideo Editor walkthrough.

7. Add graphics, captions, and title animations

Once the structure and pacing are working, add the text and graphics that help viewers follow the video.

For example:

Add short section titles for “Setup,” “Brewing,” and “Cleaning.” Fade each title in as its section begins and out before it covers an important demonstration. Add readable subtitles throughout, keeping them clear of the product and the presenter’s hands.

You can describe more specific animation:

When the presenter mentions the fill line, add a small arrow pointing to it. Have the arrow appear with the explanation and disappear before the next shot.

ChatGPT can use invideo Editor’s title and keyframe controls to build these elements. Their placement, timing, and animation remain editable.

For a custom effect or transition, provide a reference and explain the part you want to recreate. You can refine the resulting controls manually or ask ChatGPT for adjustments, including additional controls when you need them.

8. Review the complete edit and apply final changes

Watch the video from beginning to end with the visuals, sound, captions, and graphics in place.

Then give ChatGPT focused feedback:

Keep the current shot order. Hold the title for another second, and lower the music slightly during the final instruction.

If you want to try a different opening or a shorter version, ask it to duplicate the edit onto another timeline first. You can compare the alternatives while keeping the version you already approved.

Watch Luuk Alleman review playback and make a manual trim in invideo Editor.

9. Export the finished video

When you’re happy with the edit, identify the timeline and ask ChatGPT to export it as a finished video.

Export the approved “Coffee Maker Walkthrough” timeline as a finished video.

You can also export an editable project to continue working in Premiere Pro, Final Cut Pro, or DaVinci Resolve. Specify both the timeline and the application:

Export the “Coffee Maker Walkthrough” timeline as an editable project for DaVinci Resolve.

See the rendered-video and editable-project options in Devesh Sharma’s export demonstration.

Open the exported video and watch it once before sharing it. Check the beginning, ending, sound, and any text that changed during the final review.

FAQs

Can ChatGPT edit my existing video project?

Yes. Through invideo Editor’s MCP connection, ChatGPT can open an existing project, read its timeline, and continue editing. Identify the project and timeline, then explain what you want changed and anything that should remain untouched.

Is editing through ChatGPT free in invideo Editor?

invideo Editor is free to use for editing, including through MCP. Ordinary editing tasks use your ChatGPT allowance without consuming invideo credits. Generating images, video, or audio uses invideo credits separately.

Can I make manual changes after ChatGPT edits my video?

Yes. ChatGPT’s work appears on the same editable timeline you use in invideo Editor. You can trim clips, change text, adjust animation, or work on sound manually, then continue directing ChatGPT from the updated edit.

Do I need to finish the entire video in one ChatGPT conversation?

No, but staying in the same conversation helps preserve the decisions behind the edit. A new conversation can read the current timeline, but it won’t automatically know the earlier discussion. Before switching, ask for a handover document containing the brief, approved decisions, and remaining work.

Do I need one detailed prompt for the whole video?

No. Explain the overall video first, then direct the work in stages: the rough cut, revisions, supporting visuals, sound, graphics, and export. Reviewing each stage lets you make creative decisions as the edit develops.