Shazi Video Gen

Changelog

Stay up to date with the latest improvements and new features

v5.7.0 Latest
July 05, 2026
  • Migrated the AI33 Pro API to the latest **v3** version for improved performance and compatibility.
  • Added all available AI33 Pro voices.
  • Implemented voice caching for AI33 Pro to reduce loading times and improve reliability.
  • Added language-based voice filtering for faster and easier voice selection.
v5.6.0
July 01, 2026
  • Replaced **Llama 3.1 8B Instant** with **OpenAI GPT-OSS 20B**.
  • Replaced **Llama 3.3 70B Versatile** with **OpenAI GPT-OSS 120B**.
  • Fixed an issue where videos generated with the Subtitle Generator could become silent.
  • Fixed an issue where subtitles were incorrectly added to the base video.
  • Fixed subtitle timing and synchronization issues.
  • Rebuilt the subtitle generation pipeline from scratch for better accuracy and reliability.
  • Successfully tested the new subtitle pipeline on both 5-minute and 67-minute videos.
v5.5.0
May 09, 2026
  • 🖼️ Added Image-to-Video support for creating videos from AI or static images
  • 🎙️ Added Musatts Voice Clone Pro support for more realistic human-like voiceovers
  • ✍️ Improved script consistency for smarter storytelling and better flow
  • 🎬 Enhanced overall video generation quality and smoother outputs
  • ⚡ Optimized workflows and improved processing performance
  • ✅ Fixed subtitle mismatch issue for more accurate and better synchronized subtitles
  • 🕒 Improved subtitle timing stability during video playback
  • 🛠️ Minor UI/UX improvements and general bug fixes
  • 🚀 Overall stability and performance improvements across the platform
v5.0.0
April 22, 2026
  • 🧠 AI Brain Upgrade
  • Introduced Script Bible Engine
  • Reads full scripts intelligently
  • Extracts style, tone, and characters
  • Generates highly relevant (up to 99%) visuals for videos
  • Smarter scene understanding → better storytelling output
  • 🎧 Audio Improvements
  • Fixed all major Fish Audio errors
  • Added support for Fish Audio S2 model
  • Better voice quality
  • More stable generation
  • 🖼️ New Image Models (Runware + Together AI)
  • Added a powerful collection of latest image generation models:
  • Flux Series:
  • Flux Dev
  • Flux Schnell
  • FLUX.1 Schnell
  • FLUX.2 Dev
  • FLUX.2 Pro
  • FLUX.2 Klein 4B
  • FLUX.2 Klein 4B Base
  • FLUX.2 Klein 9B
  • FLUX.2 Klein 9B Base
  • FLUX.2 Klein 9B KV
  • HiDream Models:
  • HiDream-i1 Fast
  • HiDream-i1 Dev
  • Z & Turbo Models:
  • Z-Image
  • Z-Image Turbo
  • TwinFlow Z-Image Turbo
  • Other Advanced Models:
  • Juggernaut Lightning Flux
  • Imagen 4 Fast
  • GLM-Image
  • Kling Image 3.0
  • Seedream 4.0
  • Nano Banana
  • Grok Imagine Image Pro
  • ⚡ Performance & Stability
  • Improved overall generation speed
  • Optimized processing pipeline
  • Fixed multiple minor bugs & stability issues
v4.9.0
April 04, 2026
  • 🍎 Fixed FFmpeg issues on macOS
  • ⚙️ Improved macOS compatibility across different versions
  • 📜 Fixed Load Script issue (was showing success but not actually loading)
  • 💾 Fixed Custom Instruction saving issue (Windows + macOS)
  • 👀 Fixed preview issue for stock images & stock videos
  • 🔊 Fixed Fish audio Api issues on macOS
  • 🎬 Improved video quality on macOS
  • 🖼️ Fixed Replicate API issue (square images → now proper landscape)
  • 🛠️ Minor bugs and glitches fixed
v4.8.0
April 02, 2026
  • Fixed issue with square image replication
  • Fixed missing preview during generation
  • Resolved issue where videos with 300+ scenes caused problems on low-end PCs
  • Added reload and resume features for all media types (AI images, stock, and mixed content)
  • fixed Mac os Minor Issues
v4.7.0
March 28, 2026
  • Added mixed media percentage controls—manually set AI Images, Stock Videos, and Stock Photos ratio with interactive sliders
  • Added image style and model selection for Mixed Content mode
  • Fixed video preview in Review Scenes showing an error—now displays thumbnail with play overlay
  • Fixed caption highlight word timing delay—uses real word-level timestamps instead of weighted estimation
  • Fixed caption highlight drift in Shazi Renderer's native renderer
  • Fixed short stock videos freezing mid-scene—now loops seamlessly
  • Improved short video handling with Shazi Renderer's stream_loop and loop fallback
  • Added auto-transcription fallback when word timestamps are empty
  • Added mix percentage save/restore in project settings
  • Updated review page to reflect custom mix percentages
  • fixed Video Quality Issue On MacOS
v4.6.0
March 26, 2026
  • Resolved the issue where stock videos were getting stuck, along with fixing a few minor problems in mixed content creation.
v4.5.0
March 19, 2026
  • Added 100+ new image styles
  • Introduced stock videos integration
  • New mix content creation feature
  • Added search & replace for videos
  • Style search functionality for faster discovery
  • Missing styles preview now available
  • Generate sample previews for custom styles
  • Enhanced media sync pipeline
  • Fully upgraded modern UI — now control everything from one place
  • Resolved minor bugs and stability issues
  • Fixed Video Quality Issue in Mac
v4.0.1
February 19, 2026
  • Fixed an issue where projects with 100+ scenes were causing errors in the latest Shazi Render.
v4.0.0
February 15, 2026
  • Added option to disable Ken Burns effect
  • Added option to disable Heartbeat animation
  • Added option to disable Transitions
  • Introduced Multilingual Support (Beta)
  • Enhanced Subtitler System with multilingual support
  • Performance improvements and minor enhancements
  • Improved File Naming and Path Management System
v3.9.9.5 Patch
January 21, 2026
  • Voicely Server Issue Fixed
v3.9.9.5
January 19, 2026
  • Minor Bugs Fixed Related To Paths
v3.9.9
January 18, 2026
  • Fixed image regeneration in the Image Review section after creating images using Modal Flux.
  • Resolved an issue where overlay previews were not visible. The preview button now works correctly.
  • Fixed a bug occurring when loading Audio-to-Video projects.
  • Improved Audio-to-Video stability and loading system for smoother project recovery.
  • Reduced Audio-to-Video render time, significantly improving performance.
  • Enhanced the Load Project system, making it more powerful and reliable.
  • Improved face cropping accuracy in Ken Burns animations.
  • Added a Full Render Engine with support for older GPUs, delivering faster and more efficient rendering.
  • Enhanced GPU detection to optimize render speed automatically.
  • Added TTSNinja Voice Clone API integration.
  • Prepared the tool for multilingual support—full language support coming in the next release.
v3.9.8
January 08, 2026
  • 🎬 Video & Rendering Improvements
  • Enhanced Video Quality
  • Improved rendering pipeline to deliver sharper visuals and more professional-looking video output.
  • Speed Render with Overlay (Now Live)
  • The most requested feature is finally here — faster rendering without compromising overlays or visual layers.
  • Optimized Audio-to-Video Processing
  • Fixed delays where audio-to-video conversion was taking longer than expected.
  • 🧩 New Features
  • Logo Branding Support
  • You can now add your own logo directly into videos for consistent branding.
  • Heartbeat Animation for Ken Burns
  • Added a subtle heartbeat-style animation to Ken Burns effects, ideal for long scenes to keep visuals dynamic and engaging.
  • 🛠 Fixes & Cleanups
  • Media Sync Improvements
  • Resolved minor synchronization issues between audio, visuals, and animations.
  • Removed Shazi Dev Image Generator
  • Deprecated the internal dev image generator to streamline performance and focus on production-grade tools.
  • 📜 YouTube Policy Compliance Upgrade
  • Complete System Logic Recraft
  • The rendering, logging, and generation system has been fully redesigned to align with YouTube’s content policies.
  • Reuse & Authenticity Safe
  • If your script is original, well-researched, and written by you, videos generated with Shazi are fully compliant with YouTube policies (including reuse and unauthentic content guidelines).
  • 🌍 Platform & AI Enhancements
  • Multilingual System Preparation
  • Backend is now ready for multi-language support — coming very soon.
  • Enhanced Prompt Intelligence
  • Improved prompt-handling system for better content understanding, structure, and output quality.
v3.9.7
December 26, 2025
  • New Shazi Render Engine: Introduced a next-generation render engine delivering up to 5× faster rendering speeds.
  • Ken Burns Effect: Added cinematic pan & zoom (Ken Burns) effect powered by the latest Shazi Render Engine for more dynamic visuals.
  • 41+ Professional Transitions: Added 41+ high-quality video transitions for smooth and cinematic scene changes.
  • Choose a random transition Or select a specific transition from the list
  • Self-Hosted Flux Server API: Added support for modal-hosted Flux Server API, enabling faster and more flexible image generation setups.
  • Together AI Model Changes And Also Update the models to show cost
  • New cost-efficient options available at 1,000 images per $1 by Togather Ai
  • UI Reorganization: Fully redesigned interface for a cleaner layout, smoother workflow, and improved accessibility.
  • Enhanced Video Quality: Improved rendered output for sharper visuals and more professional-grade video quality.
v3.9.6
December 20, 2025
  • 🚀 New Features
  • Expanded Voiceover Library: Added GenAI Pro (genaipro.vn), AI33 Pro (ai33.pro), and TTS Ninja (ttsninja.com) as new voiceover providers.
  • New Admin Panel & Web App: Introduced a powerful web application allowing users to manage licenses, track devices, view changelogs, and access latest downloads independently.
  • ⚡ Improvements
  • Media Sync Quality: Added selectable quality presets (720p, 1080p, 2K, 4K) for Media Sync.
  • Apple Silicon Support: Optimized the application for native performance on macOS Apple Silicon chips (M1/M2/M3).
  • Responsive Dashboard: Completely revamped the Main Dashboard UI to be fully responsive on all devices.
  • 🐛 Bug Fixes
  • Audio-to-Video Engine: Fixed critical word-cutting and video duration discrepancies in the final output.
  • Subtitle Sync: Resolved audio trimming and text truncation issues in subtitle generation.
  • Media Sync Sorting: Fixed the logic for image sorting order in Media Sync.
v3.9.5 Patch 1
December 10, 2025
  • Many minor bugs removed and overall app performance optimized
  • Custom style reinforced for more accurate and stable outputs
  • Image size bug fixed – images will no longer generate in square format unexpectedly
  • Raw data issue fixed – previously, selecting only audio generated both image + audio. Now it works correctly
v3.9.5
December 08, 2025
  • Custom Image Upload – Now you can upload your own images to replace scenes instead of using AI-generated or stock visuals.
  • Improved Prompt Visibility – View scene text and image prompts separately for better control and editing.
  • Enhanced Raw Data Options – Choose to generate images only, audio/voiceovers only, or both.
  • Together AI Issue Fixed – All Together AI-related problems resolved.
  • AI Providers Reorganized – Cleaner, more structured AI image providers and model listings.
  • OpenAI GPT Image Generator Added – New model added inside AI Image Models.
  • Transition Control Feature – Add or remove transitions in your video with a single click.
  • Bulk Raw Data Generation – Now you can generate raw data in bulk mode.
  • HuggingFace Flux-Free Model Fix – All errors resolved for smooth generation.
  • Quality Presets Added – Render in 720p, 1080p, 2K, or 4K.
  • General Bug Fixes – A bunch of minor issues have been fixed.
  • Google Gemini Updated – Migrated to the latest Gemini Gen-AI 2.5 Pro & Flash models.
  • More GROQ Models – Added more fast and efficient GROQ models.
  • App Optimization – Improved stability, faster rendering, and better performance overall.