AI CREATIVE TOOLKIT

One topic in.
One finished video out.

Story Machine is a production line: AI finds a true story, breaks it into scenes, writes the prompts, paints the images, animates the clips, narrates it, writes the SEO and renders the final file. You approve each step — the machine does the rest.

Download for macOS Download for Windows Read the guide Watch demo
0visual styles
0studio modes
0steps you can edit by hand
G-Labs Story Machine
G-Labs Story Machine mode picker: Storyteller, Storyboard, Dubbing, Editor
Opening screen — pick a studio mode
STORYTELLER — THE MOST COMPLETE MODE

Eight steps, outside in

Each step is its own page in the app: stop, fix things by hand, carry on. Every screenshot below comes from a real project — “The Final Ridge”, a documentary about the 1972 Andes plane crash.

  1. 01

    Setup

    Pick your input: a topic for the AI to research, an existing .srt subtitle file, or your own script. Then set aspect ratio, the image/video/voice models and how much runs automatically.

    • Topic
    • SRT
    • Script
    • 16:9 · 9:16
    Setup page: create a video from a topic, subtitles or a script
  2. 01b

    Pick the visual style

    Still inside Setup, the Graphics block decides how the film looks. Choose the narrative angle — with characters, without characters, or the animal world — then pick one of 46+ built-in styles. Every tile is a real sample of that preset, so you see the look before you commit.

    • 46+ presets
    • Live samples
    • Cinematic · Watercolour · Anime · Stick figure
    The Graphics block: narrative angle and the grid of 46+ visual styles, each with a sample image
  3. 02

    Story

    The AI returns four real stories around your topic with a summary, why it works and its sources. You read them and pick one — the rest of the line follows that choice.

    • Sourced
    • 4 candidates
    • Pick one
    Story page: four real story candidates found by the AI
  4. 03

    Scene planning

    The script is cut into timed beats, tagged as narration / cutaway / atmosphere, and turned into an image prompt per scene. Every line stays editable before anything is drawn.

    • Timed beats
    • Image prompts
    • Line-by-line edits
    Prompt page: timed beats and one image prompt per scene
  5. 04

    Images

    Prompts go to G-Labs Automation in bulk. The grid supports multi-select, rubber-band selection, per-scene regeneration, importing your own files — and 46+ built-in visual styles.

    • Bulk
    • 46+ styles
    • Import your own
    Images page: the generated grid of Andes illustrations
  6. 05

    Videos

    Each still becomes a moving clip: the AI writes prompts describing action and camera move only, never re-describing what the image already shows. Preview in place, redo what you don't like.

    • Image → video
    • Camera prompts
    • Preview in place
    Videos page: motion prompts and the generated clip grid
  7. 06

    Voiceover

    Hands off to G-Labs Voice Studio running fully offline: one take per sentence on the subtitle's own timing, with per-line preview and regeneration.

    • Offline
    • Per sentence
    • Redo one line
    Voice page: narration generated sentence by sentence
  8. 07

    SEO

    Generates the YouTube title, description, tags and chapters — plus a purpose-made thumbnail. Copy once, paste straight into your channel.

    • Title · Tags
    • Chapters
    • Thumbnail
    SEO page: YouTube metadata and thumbnail
  9. 08

    Render studio

    Everything lands on a waveform timeline: music ducks automatically under narration, burn-in subtitles are styleable, stills get pan/zoom. Renders run in a queue and give you the final file.

    • Timeline
    • Music ducking
    • Burn-in subtitles
    Render studio: video preview and timeline
FOUR MODES, ONE STUDIO

Not only storytelling

Same models, same render room — a different way in.

🎞NARRATIVE

Storyteller

Automate narrative and documentary videos: research · script · SRT → scene plan → images → videos → edit & render. Voiceover, music, slideshow.

★ Most complete
🎬CINEMATIC

Storyboard

Build a cinematic storyboard from a script: extract characters, locations and props → reference images → shot breakdown → consistent per-shot images.

🎙LOCALIZE

Dubbing

Translate and dub a finished film: pull or transcribe its subtitles → translate with AI → new voiceover → burn subs, mix audio and render.

✂️EDIT

Editor

Cut and assemble any video: bulk drag-in, trim, split and reorder on a waveform timeline. Auto-fit clips to subtitle cues, detected silence or a BPM grid.

INSIDE

What makes the difference

🧠

Runs on the CLI you already pay for

Talks straight to Claude CLI, Antigravity, Codex or a 9Router gateway. No vendor lock-in, and no separate API key if your CLI plan already covers it.

🔌

Three apps, one conversation

Story Machine conducts, G-Labs Automation paints the images and clips, Voice Studio narrates offline — over local webhooks on your own machine.

📐

Batch work, no scene ceiling

The script decides how many scenes there are; the app sets no ceiling. Multi-select, rubber-band, bulk file import and targeted re-runs mean big and small projects are handled the same way — never one click per shot.

🎚

Your hand on every step

Every stage can pause for edits: change a prompt, swap an image, re-read one line. The AI does the heavy lifting; the last call stays yours.

Cấu hình
Config panel: pick the LLM, image, video and voice models
One config panel for LLM, image, video and voice
Windows · macOS Claude · Antigravity · Codex 46+ visual styles Image → video Offline voiceover Burn-in subtitles Music ducking Queued rendering

Start with the guide

The whole pipeline, button by button, with real screenshots of the app.

Open the Story Machine guide