The core idea
Editors wait for you to drag clips and turn dials. VibeEdit just listens. You talk — it cuts, grades, captions, builds, and hands you an MP4. You direct. It executes.
Drop in your clips and fix them by chatting. Trim the rambling, cut to the good parts, restructure the whole thing, fix the color, add captions — all by describing it. No timeline to scrub, no panels to learn.
No footage? Describe the video you want and the AI builds every scene — titles, motion graphics, transitions, music. The same chat that edits your clips can generate a video from a single sentence.
How it works
Edit real footage or build a video from scratch — same agent, same chat window.
Upload any .mp4 or .mov to the project, then describe the edit in plain English.
Cuts on word boundaries (never mid-phoneme), auto-grades each segment, burns captions with output-timeline sync, and normalizes to −14 LUFS. One encode per segment — no quality loss from chained re-encodes.
"Tighten the cut at 0:42." "Add warm grade on the b-roll." Iterate in chat. When it looks right, hit Render.
Only on VibeEdit
Lock in a recurring AI host once, and they star — consistently — in everything you make. Same face, same voice, same vibe across every upload. It becomes your channel's identity, the way CodeBullet's character is his.
No other AI video tool can keep a character consistent. We built VibeEdit around it.
Capabilities
Every operation triggered from chat. No menus, no settings panels.
Never cuts mid-phoneme — snaps to transcript word edges with a 50ms silence pad.
Analyzes each clip's luma and saturation, applies subtle ±8% corrections per segment.
2-pass −14 LUFS on the final output — broadcast-ready, no volume jumps between clips.
Word-level output-timeline offsets, 2-word UPPERCASE chunks, burned after all overlays.
Describe a scene from scratch and the agent builds it — titles, ken-burns, transitions, music.
0.25×–4.0× speed with pitch-correct audio. Green/blue screen removal per segment.
Built for every editor
YouTube, weddings, corporate, documentaries, social shorts — the agent adapts its style, pacing, and output format to match what you actually make.
Watch a real render
Prompt → preview → render → download · all in one window
Pricing
The full editor is unlocked on every plan — 4K, no watermark, every AI tool. One credit currency powers every action. Pick your monthly volume. 7-day trial on all plans.
Not ready to sign up?
Common questions
Upload any .mp4 or .mov, then describe the edit in chat. 'Trim the filler words, warm grade, burn captions' is enough. The agent transcribes, cuts on word boundaries, grades, syncs captions, and normalizes loudness — no timeline needed.
Trim to word boundaries, auto color-grade each segment, burn sync'd captions, normalize loudness (−14 LUFS), chroma key, speed ramp, mix audio tracks, create motion-graphic scenes from a text description, and layer overlays on top of real footage.
Yes. Describe a scene — '30s finance hook, black and neon green, animated counter' — and the agent builds every scene as motion graphics. Same chat, same project, same export.
Yes — h.264 with normalized audio. Drag it straight into your CMS or publishing tool.
The agent samples each clip's luma and saturation, then applies subtle ±8% corrections so segments look balanced. Override with 'warm cinematic' or 'skip grade' in chat.
They generate raw pixels you can't edit. VibeEdit edits real footage and builds compositions that render deterministically — same prompt, same MP4, every time. You can also refine anything by chat. And only VibeEdit gives you a recurring AI host who stars consistently across every video.
Generate an AI host once, lock them in, and they appear — same face, same voice, same vibe — in every video you make. It becomes your channel's identity (think CodeBullet's character). The AI can also pose and re-express your locked host for new scenes. No other AI video tool keeps a character consistent.
No. Everything renders in the cloud, right in your browser. Paid plans can add an optional local worker for faster renders.
Tell it. 'Tighten the cut at 0:42.' 'Drop the white flash.' Shift-click any preview frame to edit at that timestamp.
Yes. Drop images, video, and audio into the project. Brand kits (Studio plan) apply your logo, colors, and host identity across every new project.
Yes. Drop any footage into the project, say 'add captions', and the agent transcribes via Whisper, generates word-level offsets, styles them (2-word chunks, bold white pill), and burns them permanently into the export. Bring your own OpenAI key for transcription.
Edit your footage or build from scratch — and let your recurring host star in all of it. No timeline, no tools. Just talk.
Make your first video — free