I work on Valmera, an AI video editor with an MCP integration. Here is the editing-brief and review workflow we use to turn an open-ended request into something a creator can actually evaluate.
This guide is maintained by Valmera. It turns a broad request into an edit you can review, whether you work in Studio or through an authorized MCP client.
Start with the finished video
Give your assistant four things:
- Audience and outcome: who is watching and what they should understand.
- Source: the project or recording, plus a phrase or timestamp when useful.
- Treatment: pacing, framing, captions, music, and the intended destination.
- Protected material: statements, moments, and qualifications that must survive.
For example:
Make a concise tutorial for a first-time user. Start with the successful result, then show the steps from this recording. Remove loading waits and repeated attempts, keep every required setting, and make the interface readable on a laptop. Preserve the final explanation of how to check success.
The brief gives you a review standard: can a first-time viewer follow the steps? It is more useful than asking whether the edit looks impressive.
Six starting points
| Your recording | Ask for | Check before approving |
|---|---|---|
| A founder introduction | Remove abandoned takes and long waiting gaps while keeping natural delivery. | No clipped words; the closing invitation survives. |
| A landscape interview | Make a portrait version while keeping the relevant speaker and reactions visible. | Watch subject movement throughout each shot. |
| A webinar | Select complete practical answers that work without the full session. | Each clip retains its premise and qualifications. |
| A screen recording | Turn a single task into a clear lesson with readable controls. | Every required action remains understandable. |
| Creator product footage | Connect a real problem to a matching demonstration and invitation. | Spoken claims match what the viewer sees. |
| Travel clips | Build a small story with an arrival, experience, and ending. | Distinct shots and natural sound serve the story. |
Full examples, prompts, and review checklists:
- Remove silence naturally
- Turn horizontal footage into vertical video
- Turn a webinar into social clips
- Edit a screen recording into a tutorial
- Edit a UGC product video
- Make a travel story from phone clips
Review one editorial decision at a time
Get the story right before polishing its styling. First choose the sections that belong. Then tighten pacing. Next review framing and captions. Finally listen to the mix and inspect the export. This sequence makes feedback easier to understand and reduces repeated work on sections you later remove.
Use a revision that identifies the problem, its location, and what should stay:
In the current preview, the explanation beginning “[exact phrase]” feels rushed. Restore a little space before it. Keep the remaining cuts and caption style unchanged.
If you mention a timestamp, distinguish the original recording from the edited preview. Earlier cuts move later moments.
Use any planning assistant; connect editing deliberately
You can develop a brief in ChatGPT, Claude, Gemini, Perplexity, Copilot, Grok, or another conversational assistant. The useful output is a clear brief, not a claim that the recording has already been edited.
For direct tool execution, use the setup guide for your client and authorize your Valmera account. An assistant without that connection can still help plan the edit; paste the result into Studio.
Finish with a real deliverable
Review the rendered result with sound. Check names and numbers in captions, cuts between words, subject framing, and the final invitation. Follow the current export workflow, wait for completion, and open the resulting video before publishing it.
Top comments (0)