No. Your video never leaves the device. All frame and audio analysis runs on-device. Cloud AIs (Gemini / OpenAI / Anthropic) are controllers only — they drive the editor as text through the in-app MCP server and never receive your clips or frames.
No. There's a free, no-key, on-device path: ML Kit + MediaPipe vision for keep/remove analysis, an optional on-device LLM brain to drive the editor, a Local silence detector for audio, and free Pollinations.ai image generation. The app is fully usable with zero configuration. Cloud providers are bring-your-own-key and stored encrypted on-device if you choose to use them.
Select a clip and type the instruction. On-device vision (COCO object detection, face detection, scene labeling) scans the clip — frames are sampled at about 3 fps and a match extends a few frames either side, so it's cheap — then the clip is split at the matched boundaries and the matched pieces are deleted. To target a specific object rather than any phone, scrub to a frame showing it and say "this is my phone"; it matches that instance by image similarity.
Cutting ("cut/remove/trim the frames with X") shortens the clip — it splits and deletes the matched pieces. Erasing ("remove X but keep it natural / keep the length") keeps the clip the same length and repaints the object out using Leonardo.ai inpainting (bring your own key), producing grouped replacement segments.
Select range (dashed box) lets you drag a rectangle over the timeline to select every clip whose time span it touches, across all tracks. Ripple closes gaps: it pulls clips left to remove empty space among the selected clips, or among all clips if nothing is selected, keeping every track in sync.
Yes. While open, the app runs a small token-gated MCP server, so external AI tools (or the in-app assistant) can read the timeline and apply edits. An optional end-to-end-encrypted Cloudflare relay makes it reachable from anywhere without port-forwarding.
A real mp4, rendered with Media3 Transformer: your cuts, every composited video track, clip positions, per-clip filters and transforms, project crop/aspect, background mattes, caption overlays, and the full audio mix (volume / pan / normalize / mute / opacity). It saves to your gallery. Export runs in the background with a progress notification and can be cancelled.
Yes. Guillotine has phone, tablet, and Chromebook layouts, with keyboard shortcuts and mouse + Ctrl-scroll zoom. It's a first-class large-screen app, not just a phone port.
Regular Transcribe creates timed text clips that appear and disappear with the spoken words. Animated transcribe goes further: it splits each word into syllables on separate tracks with scale keyframes so each syllable grows from small to full size as it's spoken — a kinetic typography effect. Ask the AI for "animated captions," "kinetic text," or "per-syllable animation."
You can teach the AI a named editing method — say "create a tool called comedy zoom that does X, Y, Z" — and later invoke it on any clip with "do comedy zoom on this clip." The AI saves the step-by-step instructions and replays them using the editor's built-in tools. Use list tools to see your saved methods and delete tool to remove one.
Yes. Tell the AI "record what I do" (or "watch me edit this"), then edit the clip by hand — split, trim, delete, add keyframes, change filters, whatever you like. Every action is captured. When you're done, say "save that as X" to turn the recorded steps into a user-defined tool. You can add written caveats (e.g. "adapt timings to clip length") so the AI generalizes the method for other clips.
Yes. Open Settings → Advanced and tap Export settings to save your AI configuration (provider, API keys, models, speech/agent model paths, cache size) as a JSON file. Import settings restores them — useful for migrating to a new device or sharing a setup.
Twelve: opacity, scale, rotation, offset X, offset Y, brightness, contrast, saturation, hue, sepia, volume, and pan. Each keyframe supports per-point cubic-bezier easing with draggable handles.
Press the ? (help) button in the toolbar or top-right corner and open the icon key — it lists every icon button and what it does. The Tutorial and FAQ are in the menu (the app icon, top-left).