Stay updated with the latest features and improvements to Prologue AI.
ByteDance Seedance 2.0 — for all Business clients.
We added Seedance 2.0 across text-to-video, image-to-video, and reference-to-video (multi-image, optional video and audio guides). Expect cinematic motion, optional native audio, and the same O3-style element workflow for reference mode in Studio and the image-to-video playground.
Seedance 2.0 Text → VideoText→Videos+Animations
- Generate from prompt only—resolution 480p or 720p, duration auto or 4–15 seconds, aspect presets or auto.
- Optional generated audio (same provider cost either way); credit estimates include per-second and token-style usage.
Seedance 2.0 Image → VideoImage → Video
- Animate a start frame; optional end frame for a guided first-to-last transition.
- Matches FAL limits for image inputs; parameters mirror the public API (resolution, duration, aspect ratio, audio toggle).
Seedance 2.0 Reference → VideoImage → Video
- Combine up to nine reference images plus optional short reference videos and optional audio—refer to them as @Image1, @Video1, @Audio1 in the prompt.
- Studio passes project elements into the same flow as Kling O3 reference; when you attach reference videos, the per-second portion of pricing is discounted on the provider side—check the estimate before you run.
Multimodal Studio (beta) — the biggest Prologue drop in a year
Meet Multimodal Studio: one place where your script, storyboard, canvas, and every major generator finally work as one pipeline. Business workspaces get early access while we polish the edges—bring a draft, get structured shots with timed prompts, understand what is in frame, and launch quick actions without losing context. This is the workflow we have been building toward: pre-production through generation, connected.
One tab, every mediumCreate → Multimodal Studio
- Switch between image, video, audio, and 3D models inside the same Studio—no more hopping between separate playgrounds with lost prompts and attachments.
- Early access is rolling out on Business (and selected partner orgs) first; tell us what breaks and what you want next.
- Beta means we are still hardening performance and odd script formats; core flows—canvas, filmstrip, script ingest, shot list—are ready for real projects.
Canvas: scene detection & quick-action editsCanvas & workbench
- Scene detection finds the important pieces in a frame—people, products, environment—so you edit by intent, not by guessing the whole prompt.
- You get a readable scene description; pair it with per-object actions such as replace, remove, continuity into the next shot, reframe, extend, or upscale.
- Paint a mask, describe the fix (inpaint), queue several object-level tweaks, then apply them together in one pass into your chosen model.
- From any storyboard or history card, open quick actions: cast (character, location, prop), stage action, scene transitions, cutaways, production design, shot variations, time shift, narrative map, conversation coverage, generate world, camera angles, or jump straight to script upload.
Shots, takes & story orderFilmstrip & timeline
- Scenes and shots live in a filmstrip so you always see where you are in the story.
- Reorder scenes, add shots, and seed a shot from an upload, a library pick, or something you placed on the canvas.
- Assign a generation to the active shot; when a version wins, promote it to a named take and keep iterating without losing the runner-ups.
- Undo recent storyboard edits when you want to rewind an experiment.
Script ingestion & element detectionScript view
- Upload a screenplay or treatment (PDF or text); we parse it and spin up scenes, shots, and starter elements in your project (credit-based, like other AI steps).
- Preview the Fountain-style result before you commit; pick a single flowing column (scenario) or split video vs audio columns (commercial-style), whichever matches how you work.
- Heavy files upload through project storage first so you are not stuck under generic web upload limits.
- Characters, locations, and key props surface as elements you can cast into later generations so design stays consistent shot to shot.
Element creationElements
- When you need the same character, location, or prop to read clearly across many shots, you define it once as an element: name it, choose a type (character, scene, costume, prop, and more), and attach reference images—from uploads, generations, or your library—so the model always has a shared visual anchor instead of guessing from scratch every time.
- Build richer bibles with multi-angle references (turnarounds), optional voice for characters, and—for locations—paths toward world-style assets so environments feel continuous, not one-off sets.
- Save elements to the project and cast them from the canvas and script workflow; edits stay centralized, so when you refine an element, everything that depends on it can move forward together.
Shot list, multi-prompt beats & suggest editsScript & shot list
- Open a structured shot list: each shot shows a master video prompt plus a temporal multi-prompt sequence—timestamped beats so motion, performance, and dialogue stay aligned.
- Copy one beat or the entire breakdown for hand-off to an editor or another tool.
- Run suggest edits on the whole script or only a highlighted passage; AI proposals appear in the editor with credits shown up front.
- Review suggestions in context, then regenerate frames or video when you are happy with the words.
O3 Pro models & Stems Divide
Full O3 Pro lineup for text-to-video, image-to-video, reference-to-video, edit, and next-shot workflows and New stem separation with up to 6 outputs (Stems Divide)
O3 Pro T2VText→Videos+Animations
- Cinematic text-to-video with fluid motion
- Native audio generation (Chinese and English)
- Professional-grade outputs with precise motion control
O3 Pro I2VImage → Video
- Top-tier image-to-video with cinematic visuals
- Fluid motion and native audio generation
- Flexible duration and resolution options
O3 Pro Reference-to-VideoImage → Video
- Transform images into video with multi-reference support
- Consistent character identity and object details
- Complex scenes with multiple reference elements
O3 Pro EditVideo to Video Utilities
- Advanced video editing with natural language
- Character replacement while preserving motion
- Scene and environment transformations
- Multi-reference editing with up to 4 elements
O3 Pro Next ShotVideo to Video Utilities
- Sequential shot generation with consistent motion style
- Cinematic style preservation across new scenes
- Multi-shot narratives with consistent visual language
- Character continuity with up to 4 tracked references
Stems DivideCreate & edit audio → Music & Sound Effects
- Separate a full mix into up to 6 stems: vocals, drums, bass, other, guitar, piano
- Choose model (e.g. htdemucs_6s) and output format (WAV/MP3)
- Charged per second of audio processed
Vocals
Loading...
0:00 / 0:00
Drums
Loading...
0:00 / 0:00
Bass
Loading...
0:00 / 0:00
Other
Loading...
0:00 / 0:00
Guitar
Loading...
0:00 / 0:00
Piano
Loading...
0:00 / 0:00
Seedance 1.5 Pro (T2V & I2V)
Upgraded Seedance models to v1.5 with native audio generation, longer durations (4–12 seconds), and flexible resolution (480p, 720p, 1080p).
Seedance 1.5 T2V ProText→Videos+Animations
- High-quality video from text with optional AI-generated audio
- Durations from 4 to 12 seconds
- 480p, 720p, or 1080p output
Seedance 1.5 I2V ProImage → Video
- Animate images with optional AI-generated audio
- Optional end frame for start-to-end transitions
- 4–12 second clips at 480p, 720p, or 1080p
Business plan now live
The Business plan is now available with collaboration features, editor access management, and more.
Business plan
- Collaboration features and editor access management
- Project comments and library links
- Team members can request editor access; admins approve or reject
Project comments & library links
- Comment on projects at the project level
- Viewers can create comments and resolve threads
- Attach library items directly to comments for clearer feedback
Editor access requests
- Team members can request editor access on projects
- Admins can approve or reject requests from the project
Higher upload limits & long-running jobs
Larger file uploads and clearer handling for slow jobs like 4K video and 3D.
Higher upload limits
- Video uploads up to 500 MB
- Audio uploads up to 200 MB
- Other limits increased for smoother workflows
Long-running generations
- Get a request ID immediately for slow jobs (e.g. 4K video, 3D)
- Check job status until completion
- Credits are only deducted when the job succeeds — no charge on failure
StemGen Audio Separate
New model to separate audio into stems — isolate vocals, drums, bass, and other elements from a single track.
StemGen Audio SeparateVideo to Audio Tools
- Separate a full mix into individual stems
- Isolate vocals, drums, bass, and other instruments
- Use stems for remixes, covers, or clean versions
Original
Loading...
0:00 / 0:00
Target (isolated)
Loading...
0:00 / 0:00
Residual (rest)
Loading...
0:00 / 0:00
New enhance button logic
The prompt enhance button now works more reliably and fits better into your workflow.
Enhance button
- Improved logic for when and how the enhance button runs
- Clearer feedback and fewer edge cases
- Better integration with prompt boxes across playgrounds
Inpaint Tool Now Available with Image-1.5 Edit
Precisely edit specific areas of your images with our new Inpaint tool. Select the exact regions you want to modify using an intuitive mask editor, then let AI seamlessly regenerate those areas based on your prompt.
Inpaint ToolImage → Image - Edit
- Upload an image to the Image → Image - Edit playground
- Click on the uploaded image to view it in fullscreen
- Click the 'Edit Inpaint Area' button to open the mask editor
- Use the brush tool to paint areas you want to edit, or use the eraser to remove parts of the mask
- Use the rectangle tool for quick rectangular selections
- Adjust brush size and opacity for precise control
- Once your mask is ready, enter a prompt describing what you want in the masked area
- Generate to see AI seamlessly fill the selected regions with your desired content
BeforeAfter
Prologue Intent Now Available
Prologue Intent bridges the gap between human vision and AI execution. It translates your audio, video, and visual references into precise scripts and high-fidelity prompts that capture your true creative intent.
Prologue Intent
- Translate raw audio, video, and visual references into precise scripts
- Generate high-fidelity prompts that capture your true creative intent
- Turn a mood board and a voice memo into AI-ready creative direction.
- Align image references and sound cues into a single narrative intent.
- Reduce prompt iteration by establishing intent upfront.
- Set creative direction once, then generate consistently.
BeforeAfter
Full Platform Redesign
We've completed a comprehensive redesign of Prologue AI, bringing you a cleaner interface, improved workflows, and better performance across the platform. We value your feedback and would love to hear your thoughts on the new design.
Enhanced User Interface
- Cleaner, more intuitive design
- Improved navigation and user experience
- Better Workflows
GPT-Image-1.5 Edit Available
GPT-Image-1.5 Edit is now available for high-fidelity image editing
GPT-Image-1.5 EditImage → Image - Edit
- High-fidelity edits with strong prompt adherence
- Choose quality (low / medium / high)
- Choose image size (auto, 1024×1024, 1536×1024, 1024×1536)
- Generate up to 4 images per request
BeforeAfter
Kling 2.6 Pro Available
Top-tier image-to-video generation with cinematic visuals, fluid motion, and native audio generation. Kling 2.6 Pro delivers professional-grade cinematic outputs with precise motion control. Supports English voice output.
Kling 2.6 ProImage → Video
- Top-tier image-to-video with cinematic visuals and fluid motion
- Native audio generation supporting Chinese and English voice output
- Professional-grade cinematic outputs with precise motion control
- 5s or 10s duration options with flexible audio on/off pricing
Seedream 4.5 Models Released
Major expansion of video-to-video utilities, image-to-video capabilities, and video-to-audio tools with professional-grade models.
Seedream 4.5Text → Image
- Ultra-realistic image generation
- Unified architecture for generation and editing
- Fast 2K image generation in seconds
- 4K support for production workflows
Seedream 4.5 EditingImage → Image - Edit
- Multi-reference editing with up to 10 style/identity images
- Confined edits that preserve scene structure
- Character consistency and face cloning
- Production-ready 2K and 4K outputs
BeforeAfter
WAN 25Image → Video
- High-quality character animations with motion control
- Product showcase videos with smooth transitions
- Cinematic sequences from still images
- Social media content with optional background music
Seedance I2V ProImage → Video
- Animate headshots with natural motion
- Bring product images to life
- Add environmental effects (trees swaying, water)
MiniMax I2V-ProImage → Video
- Social clips (portrait → TikTok-style video)
- Add cinematic moves to static art
- Short ad creatives from stills
ExtensionProVideo to Video Utilities
- Extend short clips with matching motion and context
- Create natural transitions at the tail of a shot
- Finish cuts for social or ad content
ReframeVideo to Video Utilities
- Convert landscape videos to portrait for social media (TikTok, Instagram Reels)
- Transform portrait videos to landscape for YouTube or widescreen displays
- Adapt existing content to different aspect ratios without manual cropping
- Maintain visual quality while changing video format for different platforms
O1 Next ShotVideo to Video Utilities
- Sequential shot generation maintaining motion style and camera language
- Cinematic style preservation across new scenes
- Multi-shot narratives with consistent visual language
- Character continuity across generated shots with up to 4 tracked references
- Production extensions with matching cinematic feel
Auto SubtitleVideo to Video Utilities
- Social media content: TikTok, Instagram Reels, YouTube Shorts with animated subtitles
- Accessibility: Add captions for hearing-impaired viewers
- Multi-language content: Transcribe videos in 10+ languages automatically
- Professional videos: Add branded subtitles with custom fonts and colors
- Educational content: Create clear, readable subtitles for tutorials
- Marketing videos: Enhance engagement with karaoke-style word highlighting
SyncWaveVideo to Audio Tools
- Craft background music for video
- Add creative audio layers that follow visuals
- Test different moods/genres on the same clip
FoleySFXVideo to Audio Tools
- Add realistic foley (footsteps, doors, clinks)
- Create ASMR-like soundscapes
- Replace or enhance audio tracks without recording
O1 Models Available
Create or modify images and videos while maintaining consistency and continuity with our new O1 models.
Kling O1 EditVideo to Video Utilities
- Advanced video editing with natural language instructions
- Character replacement while preserving motion
- Scene environment transformations
- Multi-reference editing with up to 4 elements/images
O1 Reference-to-VideoImage → Video
- Transform images into consistent, high-quality video scenes
- Stable character identity and object details
- Multi-reference support for complex scenes
FLUX.2 Models Available
Next-generation Flux models with advanced prompt understanding and professional-grade image generation and editing capabilities.
Flux2 FlexText → Image
- High-quality image generation with advanced prompt understanding
- Support for @{field} syntax for referencing uploaded images
- Professional-grade visuals up to 2K
FluxFlex-2-EditingImage → Image - Edit
- Enhanced image editing with multi-reference support
- Up to 8 images via API (9MP total limit)
- Color matching with hex codes or image references
- Natural language editing instructions
BeforeAfter
Nano-Banana Pro Model Launch
Google's state-of-the-art Nano-Banana Pro model delivers exceptional prompt adherence and sophisticated photo editing capabilities.
Nano-Banana ProText → Image
- Ultra-realistic image generation with exceptional prompt adherence
- Sophisticated photo editing using conversational text prompts
- Multi-image editing support with up to 10 reference images
- 4K output support for production workflows
- Advanced composition control
Nano-Banana Pro EditImage → Image - Edit
- Conversational text prompts for image editing
- Multi-image editing support
- Exceptional prompt adherence
- 4K output support
- Advanced composition control
BeforeAfter
SAM-3 Object Detection and Segmentation
Advanced 2-layer object detection system using Gemini 3.1 Pro for semantic understanding and SAM-3 for precise pixel-level segmentation.
MaskGen Videos (SA2VA)Video to Video Utilities
- Semantic scene understanding with Gemini 3.1 Pro
- Precise image segmentation with SAM-3
- Video segmentation with tracking support
- Multi-concept detection and masking
- Enhanced content editing capabilities
- Great for masks/segmentation/rotoscopy
- Object-centric analysis and shot planning
Veo 3.1 with Cinematic Control and Sound Integration
Google's Veo 3.1 model now supports integrated sound generation and advanced cinematic controls for professional video production.
Veo 3.1Text→Videos+Animations
- Text-to-video with native audio generation
- Cinematic control parameters for precise motion
- Enhanced video quality up to 1080p
Veo 3.1 Image-to-VideoImage → Video
- Animate still images with cinematic motion
- Reference-to-video mode for style consistency
- First-last-frame-to-video for controlled sequences
- Enhanced video quality up to 1080p
Veo 3.1 Reference-to-VideoImage → Video
- Reference-to-video mode for style consistency
- Multi-reference image support
- Controlled video sequences
Veo 3.1 First-Last-Frame-to-VideoImage → Video
- First-last-frame-to-video for controlled sequences
- Precise motion control between frames
- Cinematic transitions
Sora 2 Pro Models Available
OpenAI's Sora 2 Pro models now available for high-fidelity text-to-video and image-to-video generation.
Sora 2 T2V ProText→Videos+Animations
- High-quality video generation from text
- 720p and 1080p output support
- Enhanced visual fidelity
- Cinematic quality outputs
Sora 2 I2V ProImage → Video
- Cinematic motion from still images
- 720p and 1080p output support
- Enhanced visual fidelity
- Professional-grade video generation
Sora 2 RemixVideo to Video Utilities
- Video remix capabilities
- Transform weather/lighting in videos
- Style transfers and seasonal changes
- Object transformations and mood shifts
Seedream 4 and Seedance Models Integration
ByteDance's latest Seedream 4.0 (T2I/I2I) and Seedance (T2V/I2V) models bring unified generation and editing workflows.
Seedance T2V ProText→Videos+Animations
- High-quality video generation from text
- Expressive motion and cinematic quality
- 1080p output support
LipSync Pro Released
Professional lip-sync correction tool for fixing audio-video synchronization issues in video content.
LipSync ProVideo to Audio Tools
- Fix lip-sync issues in dubbed content
- Animate avatars to match voiceovers
- Match podcast or narration audio to video
20+ New AI Models Added
Major expansion of our model library with 20+ new models across image generation, video creation, and audio processing.
RecraftText → Image
- Creative image generation
- Style-focused outputs
HiDreamText → Image
- High-quality image generation
- Professional outputs
FluxDevText → Image
- Fast image generation
- Creative workflows
Recraft Image-to-ImageImage → Image - Edit
- Style transfer
- Image transformation
BeforeAfter
@page releaseText→Videos+Animations
- Master-quality video generation
- Professional outputs
PixVerseText→Videos+Animations
- Fast video generation
- Creative video outputs
Kling Master I2VImage → Video
- Master-quality animation
- Professional motion
PixVerse I2VImage → Video
- Fast image animation
- Creative motion