VidTL Editor is a desktop app and does not work on mobile devices. Please open this page on a desktop computer to continue.
Changes to resolution and frame rate will affect all clips. Audio settings apply to new exports.
Your browser paused video playback to save memory because the editor was inactive for a while.
Don't worry - your work is safe! Click the button below to resume editing.
View and customize your hotkeys
Reduce steady background noise and static locally in your browser.
Selected video clips use their linked audio when available.
Analyzing selected audio...
Auto-set from the selected clip. Higher values allow stronger noise reduction.
Auto-set from the selected clip. Higher values can reduce leftover static.
The processed WAV is also added to the media library.
First run loads the local model, then processes audio on this device.
Let an AI agent running in your terminal (like Claude Code) edit this project for you. Create an access key, hand it to your agent, then turn on agent control below. You can watch every change live and undo anything.
You have no access key yet.
Copy your key now — it will not be shown again.
1. In your terminal, download the connector:
curl -O https://vidtl.com/app/cli/vidtl.mjs
2. Give your agent the key:
export VIDTL_API_KEY=<your access key>
3. Ask your agent to start with:
node vidtl.mjs instructions
4. Using Claude Code? Run this once per project so your requests are always logged word for word:
node vidtl.mjs install-hook
While this is on, your connected agent can see and edit this project.
Agent control is off.
Detects and removes silence using waveform analysis.
When unchecked, only processes selected clip(s)
Uses subtitle density to detect talking-head clips and skip b-roll footage
Negative = cut into audio, Positive = leave padding
Red highlights on timeline show what will be removed
Select your music clip plus the video or image clips for the montage. The editor listens for the moments that punch — vocals coming in, drum hits, drops — and cuts your clips so every transition lands on one, the way music videos are cut.
Add markers to cut by hand, or build the montage automatically
Finds the camera cuts inside your footage so long recordings can be split into separate scenes automatically.
When unchecked, only the selected clip(s) are analyzed
Higher finds more cuts, lower only finds strong cuts
Select your synced camera clips. The editor listens for who is speaking and cuts between cameras automatically. Line the clips up first with Multi Cam Smart Sync.
Shorter values switch cameras more often
When unchecked, you only hear the camera that is on screen
Made for screen recordings. Select the recording, and the editor finds the moments where something happens on screen — clicks, typing, windows opening — and zooms in on that spot automatically. Any zoom animation already on the clip is replaced.
How far to zoom in on the action
Higher reacts to smaller on-screen changes
You can adjust the zoom amount before applying
Brings your clips to the same comfortable loudness — quiet clips come up, loud ones come down, and nothing distorts. Works on the selected clips, or every audio clip when nothing is selected.
Steadies shaky handheld footage. Select the shaky clip first. The steadied video is saved to your media library — this can take a while, so it works best on short clips.
Higher handles bigger shakes but may crop in more at the edges
The original file stays in your media library
Separates the voice from music, noise, and background sound. Select the clips to clean up first. The separated audio is saved to your media library. The first use downloads the separation model, and processing takes a while for long clips.
The original audio stays in your media library
Splits the sound in two: the voice stays here, and everything else goes to a separate audio track so you can turn it down or delete it later
Lower sensitivity keeps more of the subject when the mask removes too much.
Improves edges around hair and fine details, but takes longer.
These settings are used when refined edges is enabled.
Lower keeps more of the subject. Higher removes more uncertain pixels.
Softens the transparent edge after the mask is created.
Select one or more clips on the timeline to generate subtitles with word-by-word timestamps.
Convert text to speech using different AI models. Maximum 5000 characters.
Generate original music from a text description.
Create a brand-new video clip from a text description. This is the slowest, most demanding option — try stock footage, your own uploads, and motion graphics first.
Uses your existing subtitle clips to find matching stock video and place it on the timeline.
Run custom FFmpeg commands on your media files.
Note: Input files can be any size, but output files are capped at 2GB to prevent browser issues.
Click a file to insert its name into the command
Filenames with spaces will be automatically quoted
Used if not specified in command
Tip: -ss before -i for faster/accurate seeking
Processing with FFmpeg...
Higher quality = larger file size. For ProRes, this selects different quality levels.
Available for WebM exports. Empty canvas areas stay transparent.
Use Mono if your exported video only has audio in one speaker
Each section between markers downloads as its own video — great for chapters and short clips. Not available for GIF.
One video downloads per checked size. The picture is scaled to fill each size and centered, so the edges may be trimmed. Not available for GIF.
Smaller files with a little less detail. Recommended for smooth, reliable WebM exports.
Select a platform preset to automatically configure optimal settings
Keep this tab open and the browser window visible while exporting. The window can be in the background, but minimizing it may pause the export on some browsers.
Compatible with Premiere Pro, Final Cut Pro 7, DaVinci Resolve
Clips with looks that other editors cannot recreate (effects, masks, animated graphics) will be rendered as video when exporting with media.
A browser-based video editor.
Copyright 2026 VidTL