Step-by-step guides to everything in Hiroska: timeline editing, captions, green screen, keyframes, audio mixing, AI generation, trained styles and export.
Getting started
Make your first video — From an empty project to an exported MP4: create a project, add or generate media, arrange the timeline and export in 4K. A five-minute walkthrough.
Import and organize media — Upload footage, record voiceovers, tag and search your library, and reuse assets across every project in the workspace. How the media library works.
Export your video — Everything about exporting: resolutions up to 4K, quality presets, H.264 against H.265, transparent WebM, GIF and audio-only MP3, all without a watermark.
Editing
Timeline basics: trim, split & arrange — Learn the timeline: trimming, splitting, closing gaps, magnetic snapping, multi-select, keyboard nudges and undo. The editing core of Hiroska.
Layers & picture-in-picture — Stack video lanes, turn any clip into a floating overlay, drag and resize it on the canvas, and round its corners. Multi-layer compositing explained.
Add text & titles — Step by step: add a title or a lower third, style it, place it on the frame, time it and animate it in and out. Plus when to use Graphics instead.
Generate captions automatically — Transcribe speech into captions in 30+ languages, edit every line, restyle them like any text layer, and burn them into the export. Step by step.
Remove silences & filler words — Tighten a talking-head video in two passes: scan for quiet stretches and jump-cut them, then find every filler word. You approve each cut from a checklist.
Add transitions between clips — Put cross-dissolves, directional slides, zoom punches, fades and dips between clips, set their duration, and keep effects intact through the transition.
Animate with keyframes — A walkthrough of the keyframe editor: set the first key, set the second, pick an easing per segment, and stack properties for a fly-in or a pop.
Crop, rotate & frame clips — Crop any region, rotate and flip, tilt at a free angle, and choose how a clip fills the canvas: fit, fill, or a blurred background for mismatched shapes.
Use the Ken Burns effect — Give still images cinematic motion: frame a start window and an end window, and the editor animates the zoom and pan between them. With framing tips.
Speed, reverse & freeze frames — Slow motion, speed-ups, reverse playback and freeze frames: how the playback-time tools work on the Hiroska timeline, and when to reach for each one.
Effects
Remove a green screen — Chroma key step by step: pick the key colour, tune similarity and edge blend, kill green spill, and put any background behind your subject.
Erase parts of a video — Paint any region of a clip transparent with a soft brush: remove logos, cut windows, knock out backgrounds, and refine the painted area at any time.
Blur, masks & spotlight — Hide a face or logo behind a blur, crop a clip to a feathered shape, or dim the frame around a spotlight. Which of the three tools to use, and when.
Color grade your video — Balance an exposure first, try a preset look on top, then load a .cube LUT and match your shots. Where each control lives, and what order to use them in.
Audio
Mix audio & duck music under speech — Stack music, voice and effects on audio lanes, duck the music under speech automatically, shape the fades, and clean up noisy recordings at export.
Generate music, voiceovers & sound effects — Pick the sound type, describe it, match foley to a clip, and drop the result into the mix. A walkthrough of the sound side of the Generate tab.
Clean up a recording & split a song into stems — Rescue a noisy voiceover, isolate a voice, pull the vocals or drums out of a song, extend a track to fit your edit, or swap the voice. From the Audio tab.
Clone your voice & narrate with it — Record thirty seconds once, then narrate any clip or a whole timed script in your own voice. Setting a voice up, picking a model, permission and deletion.
AI generation
Generate AI video — Generate video from text, animate stills, morph between frames and extend clips. How to choose a task, a model, the controls and the variants.
Generate & edit images — Create images from text, edit them with up to 14 reference images, remove backgrounds, and animate any result into video. The image round trip.
Upscale video & images — Run clips and stills through Topaz-class AI upscaling at 2× or 4×: sharpen soft footage, and make small AI generations good enough to export.
AI tokens, pricing & refunds — What tokens are, how each generation is priced, monthly allowances against top-up packs, and the cases where you are refunded automatically.
Transform a clip you already have — Change footage you already have: restyle a shot from a prompt, reframe it without cropping, erase something from it, apply a look, or dub it.
Train Hiroska on your own style — Train a style on a dozen of your own pictures and every generation comes out in your look. Which files to choose, what it costs, how to fix a weak style.
Dub a video or a voiceover — Translate the speech in a clip and re-sync the mouth, or dub a bare voiceover with no picture involved. Thirteen languages, the same voice and timing.
Lip-sync a clip to new audio — Reshape a speaker's mouth to match a different audio track, whether a dub, a re-record or a script change. Or make a single still portrait talk.
Turn one edit into every aspect ratio — Re-crop a finished video for 9:16, 1:1 or 16:9 and keep the subject in frame. AI reframing generates the missing sides instead of throwing the edges away.
Connect Claude, ChatGPT or Cursor to Hiroska — Add Hiroska as a connector in claude.ai or ChatGPT, or give Claude Code and Cursor an API key, set a daily token limit, and ask the assistant for a video.