User Guide

Creating Sequences

How to create and manage video sequences in OpenStory

Sequences are the core unit of work in OpenStory. Each sequence represents a video project that progresses from script to finished video.

Creating a New Sequence

  1. Navigate to Sequences from the top navigation
  2. Click Create New Sequence
  3. You'll be taken to the script editor where you can configure your sequence

Writing Your Script

The script editor accepts anything from a one-line idea to a full screenplay:

  • One-liner or URL — Type a brief concept like "A cat exploring a cyberpunk city at night" or paste a website URL. Use Enhance Script to expand it into a full screenplay.
  • Full screenplay — Paste a complete screenplay and click Generate Sequence directly.

The editor shows a placeholder guiding you: "A one-liner or website URL is all you need — click Enhance Script to do the rest. Or paste a full screenplay and generate directly."

Enhance Script

For short scripts (under ~1000 characters), OpenStory nudges you to enhance before generating. The Enhance Script button (bottom-right of the editor, marked with a sparkle icon) uses AI to expand your idea into a detailed screenplay with:

  • Visual descriptions
  • Camera directions
  • Scene breakdowns
  • Dialogue and action lines

Before enhancing, you choose a target video duration:

PresetDuration
15s15 seconds
30s30 seconds
1m1 minute
2m2 minutes
3m3 minutes

The enhanced script streams in real-time. You can:

  • Stop enhancement mid-stream (click Stop or press Esc / Cmd+.)
  • Undo the enhancement to restore your original script
  • Edit the enhanced script before generating

Keyboard Shortcuts

ShortcutAction
Cmd + EnterGenerate sequence
Esc or Cmd + .Stop script enhancement

Generation Settings

Click the settings button in the control bar to configure all generation parameters. Settings are persisted in localStorage so they carry across sessions.

Aspect Ratio

Choose the output aspect ratio for all generated images and videos:

RatioUse Case
16:9Landscape — standard widescreen video
9:16Portrait — vertical/mobile video (TikTok, Reels, Shorts)
1:1Square — social media posts

Analysis Model

The AI model used to analyze your script and break it into scenes. You can select multiple models to generate parallel sequences — one per model — for comparison. Available models include options from Anthropic, Google, OpenAI, xAI, Mistral, DeepSeek, and more.

When multiple models are selected, the footer shows "N sequences will be created" to confirm.

Image Model

The AI model for generating scene images. You can select multiple image models to generate variant images across different providers. Available models include:

ModelProviderNotes
Nano Banana 2GoogleFast generation and editing (default)
Nano Banana ProGoogleEnhanced realism and typography
Grok Imagine ImageGrokAesthetic generation with low censoring
FLUX.2 MaxBlack Forest LabsExceptional realism and consistency
PhotaPhotaCharacter consistency via profiles
Hunyuan Image v3TencentOpen source with strong composition
FLUX.2 DevBlack Forest LabsOpen source, 32B open weights
Qwen Image 2 ProAlibabaApache 2.0, native 2K, text rendering
HiDream I1HiDreamMIT licensed, 17B parameters
Seedream 5ByteDanceUnified generation and editing

Motion Model

The AI model for image-to-video animation. Toggle Auto-generate motion to automatically create video clips for each scene after images are generated.

ModelProviderEst. Time
LTX 2.3 ProLightricks~15s (open-source)
Veo 3.1Google~25s
Kling v3 ProKling~20s (default)
Grok Imagine VideoGrok~20s
MiniMax Hailuo 02MiniMax~15s
Seedance 1.5 ProByteDance~12s
Seedance 2ByteDance~20s (animation styles only)

Some motion models are restricted to specific style categories (e.g., Seedance 2 requires an animation style).

Music Model

Toggle Auto-generate music to automatically create a soundtrack after scene generation. Choose from:

ModelProviderMax DurationType
ElevenLabs MusicElevenLabs600sMusic (default)
MiniMax Music v2MiniMax300sMusic (supports lyrics)
ACE-Step 1.5ACE Studio240sMusic (open-source)
Lyria 2Google30sMusic
MMAudio V2MMAudio8sSFX (video-to-audio)
ElevenLabs SFXElevenLabs22sSound Effects

Pre-Generation Options

Before generating, you can optionally attach resources from your libraries:

Talent Suggestions

Click the talent icon to attach characters from your Talent Library. The AI will use these as visual references when generating character appearances, ensuring consistency with your established character designs.

Location Suggestions

Click the location icon to attach locations from your Location Library. The AI will reference these when generating scenes set in matching environments.

Both are marked as "optional" in the UI — sequences work fine without them.

Style Selection

Below the script editor, choose a visual style that defines the aesthetic of your sequence. Styles appear as a grid of tiles with preview images. Click any tile to select it, or click More to browse the full style catalog in a dialog.

Each style includes configuration for color palette, artistic direction, and rendering approach. The selected style influences:

  • Image generation prompts
  • Color palette and mood
  • Camera and lighting defaults

The Generation Pipeline

When you click Generate Sequence, OpenStory runs an automated pipeline:

  1. Script Analysis — The AI model breaks your script into individual scenes with metadata (title, duration, location, time of day, story beat)
  2. Character Extraction — Characters are identified from the script with physical descriptions, clothing, and consistency tags
  3. Location Extraction — Settings are extracted with architectural details, lighting, and ambiance
  4. Image Prompts — Visual prompts are generated for each scene incorporating style, characters, and locations
  5. Image Generation — Scene images are generated using your selected image model(s)
  6. Motion Generation — If auto-motion is enabled, each image is animated into a video clip
  7. Music Generation — If auto-music is enabled, a soundtrack is generated based on the script's mood

Progress is shown via a real-time banner with phase indicators. You can navigate away and return — the process continues in the background via durable workflows.

Regenerating a Sequence

From the Script tab of an existing sequence, you can edit the script and click Regenerate Sequence. This creates a new sequence from the modified script (the original is preserved).