Captions in 30+ languages
Whisper speech recognition with real language detection, proper scripts for Hindi, Bengali or Arabic, and translation.
Captions in 30+ languages, background and noise removal, highlights, thumbnails, stabilisation — VDOAPP runs every AI model inside your browser, on your own computer or phone. Pro editing tools included, up to 4K.
No account · Nothing to install · Works offline once models are cached
Each tool downloads its model once, then runs on your graphics card or processor. Your footage, voice and faces stay with you.
Whisper speech recognition with real language detection, proper scripts for Hindi, Bengali or Arabic, and translation.
Blur, recolour, spotlight or replace what's behind a person — no green screen. Any subject in photos.
Neural noise suppression removes fans, traffic and hiss; silence removal tightens every pause.
Scores every moment for sound, faces, motion and focus, and cuts a 15, 30 or 60-second short.
Finds your sharpest, most expressive frame, makes the subject pop and places the title away from the face.
Follows the person when you go vertical, steadies handheld shake and auto-zooms on emphasised words.
Automatic face blur for privacy, skin smoothing and face stickers that follow every face.
Nine natural US and UK English voices generated from your script, placed at the playhead.
The AI does the tedious parts. These are the tools for everything else.
Trim, split (S), markers (M), snapping to clips, markers and music beats, and 100 steps of undo.
Montage, Hero, Bullet and custom speed curves; keyframe position, scale, rotation, opacity and volume.
From dissolve and wipe to whip pan, spin, glitch and zoom punch — previewed before you apply them.
Import .cube LUTs with intensity control, plus temperature, tint, vignette, grain and film effects.
Overlays in rectangle, rounded or circle shapes, blend modes, and a green-screen key with spill clean-up.
Outline, shadow, box and neon styles, in/loop/out animations, and your own uploaded fonts.
Per-group levels, 14 built-in sound effects, voiceover recording and automatic music ducking.
Preview copies for smooth 4K, .ts and HEVC support, and frame-perfect MP4 export up to 4K at 60 fps.
Switch the canvas between 10 aspect ratios. Fit with a blurred or coloured background, fill and crop — or let Smart reframe pan the crop to follow the person, so a landscape interview becomes a vertical short without losing the speaker.
Videos, photos, music, .ts dash-cam files — your browser reads them from your disk. Nothing is uploaded.
The first time you use a tool, its model downloads and is cached. From then on it runs locally — even offline.
Save MP4 up to 4K with no watermark. Your project autosaves in your browser so you can continue later.
Tested, in-depth guides to the techniques behind professional video — and honest about what AI can and can't do.
Why captions fail in Hindi or Bengali, how to pick a model, and how to fix the last 5%.
Read guide Short-formA repeatable workflow with AI highlights, smart reframe and captions.
Read guide WorkflowClean sound, tight pauses, auto zoom, B-roll and captions — in the order that saves time.
Read guide AI effectsHow AI cut-outs work, how to film for them and five looks that hide a messy room.
Read guide Pro editingPreview copies, fast export and memory — how to edit 4K smoothly.
Read guide How it worksWhich models run in VDOAPP, what they download, and why nothing is uploaded.
Read guideTest results, creator workflows and notes from building an AI video editor that runs on your device.
WebGPU, WebAssembly and WebCodecs in plain English — and how to check your browser.
GuidesPrivacy, pricing, languages, watermarks and exports — what to check before you commit.
TipsLosing the clip, over-editing, posting too soon — how to keep dash cam footage useful.
VDOAPP downloads open AI models once from public model hosts and runs them in your browser using your graphics card (WebGPU) or processor (WebAssembly). Your media is processed on your device and never sent to us or to an AI service. Read how it works.
Yes — no account, no credits, no time limit and no watermark. Because the AI runs on your device, our costs are low; ads on our website pages cover them. There are no ads inside the editor.
More than 30, including English, Spanish, Hindi, Bengali, Urdu, Arabic, Tamil, Chinese, Japanese and Swahili, with automatic detection. Captions can be translated into 12 languages. Languages that small models handle poorly automatically use the largest model your device supports.
Most AI tools work in any recent browser. The largest caption model needs WebGPU, available in current Chrome and Edge on most computers. Speed depends on your device, and each model is downloaded once and then cached.
Yes. Make a light preview copy for smooth 4K editing (export still uses the original), open iPhone HEVC videos where your browser can decode them, and open .ts, .mts and .m2ts dash cam and camcorder files directly.
MP4 from 480p to 4K at 24, 25, 30 or 60 fps, with no watermark. When possible, fast export renders faster than real time and frame-perfect; otherwise the editor exports in real time.
No downloads, no sign-up, no uploads. Open the editor and start.
Open VDOAPP Editor