Getting Started
Everything you need to know to start editing with AI — from uploading your first footage to getting the most out of the agent.
Your First Project
Vyra is a video editor built around an AI agent. You upload your footage, and the agent helps you assemble, search, and edit — from a first rough cut to polished motion graphics.
To get started, create a new project and add your footage. Once your files are uploaded and processed, you can start editing — either manually through the UI like a traditional editor, or by talking to the agent through conversation.
Footage & Storage
Vyra has two ways to handle your footage, and it's worth understanding the difference early on.
Local Mode
When you add files directly to a project, the original stays on your device and is never uploaded. Editing happens against the local file. Fast to start, but the project only works on the device where the original file lives.
Cloud Mode
Files uploaded to your asset library are stored in the cloud. Projects using these assets can be edited from any device. If you need to edit from multiple machines, make sure all project assets are backed up.
Tip: You can back up local assets to the cloud at any time from the cloud icon in the project's top navbar. This converts them to cloud mode so you can edit from anywhere.
Regardless of how you upload, all footage is automatically processed and indexed. This means the agent can understand what's in your footage — every scene, every moment — which powers semantic search and intelligent editing.
Semantic Search
Traditional editors make you search by filename. Vyra understands your footage.
After your files are processed, you can search for moments using natural language. Instead of hunting for DSC_0042.MOV, you can search for things like:
Search results return specific clips and timestamps — not just whole files. This works within a single project or across your entire library, so you can pull the perfect moment from any video you've ever uploaded.
Setting Up Your AI
Vyra's AI agent connects through MCP (Model Context Protocol). This means you can use it from Claude, Cursor, Windsurf, or any MCP-compatible client — not just the built-in chat.
MCP Setup Guide
Step-by-step instructions for connecting Vyra to your preferred AI client.
What the Agent Can Do
The agent has full access to every feature in the editor. Anything you can do through the UI, you can ask the agent to do through conversation.
The agent can manipulate your timeline the same way you would — adding clips, rearranging them, trimming, and cleaning up. It understands multi-track layouts and can work across many items at once.
- Add, move, trim, split, and delete clips — place items at specific times, trim to exact frames, split at the playhead
- Arrange tracks and layering — move items between tracks, reorder layers, consolidate empty tracks
- Batch operations — move, edit, or delete multiple items at once. Ripple edits push later items when inserting or removing
- Slip editing — shift which portion of a source clip is visible without changing its position or duration on the timeline
- Markers — add, rename, and manage timeline markers to flag important moments
Agent vs. Doing It Yourself
The agent can do everything, but that doesn't mean it should. Some things are faster by hand — especially small, precise adjustments.
Let the agent handle
- First rough cut from raw footage
- Searching hours of footage
- Transcript-based editing
- Generating motion graphics
Either works
- Effects & color grading
- Captions & text overlays
- Smart masks & tracking
Faster to do yourself
- Fine-tuning trim points
- Nudging overlays by pixels
- Precise logo positioning
Tip: A good workflow: let the agent build the rough cut and handle the heavy lifting (searching footage, assembling, generating graphics), then fine-tune timing and positioning yourself in the UI.
Communicating with the Agent
The agent sees your editor state — your selection, playhead position, and timeline. The more context you give it, the better the results.
Be specific in your requests
Instead of “make this better,” try “increase the contrast on this clip and add a subtle vignette.” Specific requests get specific results.
The agent sees your selection and playhead
Select items on the timeline and refer to them as “the selected clip.” Move your playhead to a specific moment and say “at this point in the video...” The agent knows what you're looking at.
Highlight a range on the timeline
Drag along the ruler at the top of the timeline to select a time range. Then say “in this section, speed up the pacing” or “add a transition here.”
Ask the agent to look at what you see
The agent can capture the current frame to see exactly what's on screen. Say “can you look at what this frame looks like?” and it will capture and analyze it.
Ask the agent about the UI
Not sure how to do something? The agent has comprehensive documentation access. Ask “how do I add a mask?” or “where are the export settings?”
Reference Videos
Reference videos are one of the most powerful features in Vyra. They give the agent a visual guide for the style, pacing, and feel of what you're making.
Where to find them
In the left panel, next to the Footage tab, you'll find the References tab. You can add up to one reference video per project.
What the agent sees
The agent analyzes your reference for editing style, pacing, font choices, motion graphic style, transitions, and overall feel. It uses this as a creative guide when making decisions about your edit.
Point to specific moments
You can reference specific timestamps: “look at the transition at 0:45 in the reference” or “match the text style from the intro.” The agent can capture frames from the reference video to understand exactly what you mean.
Note: Reference videos are for inspiration and style — they're not your footage. The agent won't use clips from the reference in your edit. It only uses the reference to understand the vibe you're going for.
Motion Graphics
Motion graphics in Vyra are generated by the agent as code. This means they're fully dynamic — animated, responsive, and unique to your project. But it also means you edit them through conversation, not by dragging handles.
How it works
Describe what you want
“Create a lower third with the name John Smith and the title Creative Director” — the more detail you give, the closer the first result.
The agent generates and places it
It writes animated code and places the motion graphic on your timeline. You'll see it appear in real time.
Refine through conversation
To modify it, ask the agent — “make the text bigger” or “change the animation to slide in from the left.”
Important: You can't manually edit motion graphics by clicking on them — they're rendered from code. To make changes, ask the agent. It can modify colors, text, animations, sizing, and timing.
Your assets can be included. Images, logos, and other media from your project can be embedded directly into motion graphics. The agent can animate them, add borders and frames, or compose them into complex animated layouts.