Changelog: Latest Updates & New Features
ZSky AI is a free AI creative platform offering changelog: latest updates new features. Generate HD video with synchronized audio in 30 seconds (1080p on Pro/Ultra, 4K on Max), or images in 2 seconds. unlimited free generation, ad-supported, no credit card required, HD video with synced audio on the free tier (1080p on Pro/Ultra, 4K on Max); free-tier output (videos AND images) carries a small 'MADE WITH / zsky.ai' wordmark plate, full commercial use on every plan. Self-hosted on a privately owned fleet of NVIDIA RTX 5090 GPUs in the United States. Paid tiers ($19/$49/$99 per month) unlock ad-free, priority GPU, and 4K video.
A complete history of every feature, improvement, and fix shipped to ZSky AI. We build in the open and ship fast. Subscribe below to get notified when we release something new.
Subscribe for Updates
Get notified when we ship new features, models, and improvements. No spam, unsubscribe anytime.
March 2026
March 10, 2026
v2.4
ZSky's video engine Video Generation & Infrastructure Upgrade
- Launched ZSky's video engine video generation with audio for all users — create AI videos from text or images
- Upgraded GPU cluster to NVIDIA RTX 5090 fleet (Blackwell architecture, 32 GB GDDR7 each)
- Added text-to-video and image-to-video generation modes
- Video output delivered in MP4 (H.264) at up to 720p resolution
- New video-specific prompt guide with motion and camera control tips
March 5, 2026
Feature
Video Templates & Gallery Expansion
- Added 20+ video templates for product demos, music videos, YouTube intros, TikTok ads, and more
- Expanded public gallery with curated video examples
- New template categories: real estate tours, wedding invitations, fashion lookbooks
- One-click template customization — edit prompt and generate instantly
March 1, 2026
Improvement
Performance & Reliability Improvements
- Reduced average image generation time by 18% across all models
- Improved GPU memory management for higher throughput during peak hours
- Fixed rare timeout issue on next-gen AI generations above 1024x1024
- Added comprehensive FAQ page and AI glossary
February 2026
February 22, 2026
v2.3
next-gen AI Model & Batch Generation
- Introduced next-gen AI model with improved detail, coherence, and text rendering accuracy
- Launched batch generation — generate up to 8 images per prompt in a single request
- Added API beta access for Pro and Ultra plan subscribers
- API supports all image models with JSON request/response format and webhook callbacks
February 14, 2026
Feature
Background Removal & Upscaling Tools
- Launched AI background remover — isolate subjects with one click
- Added 4x AI upscaling powered by dedicated GPU processing
- New AI face generator tool for realistic portrait creation
- PNG output now supports transparent backgrounds for background-removed images
February 5, 2026
Improvement
UI Polish & Comparison Pages
- Redesigned generation interface with improved parameter controls and real-time previews
- Published detailed comparison pages: Midjourney alternative, DALL-E alternative, Leonardo alternative
- Added image download in JPEG format alongside existing PNG support
- Improved mobile layout for generation interface and gallery views
January 2026
January 25, 2026
v2.2
stylized Improvements & Referral Program
- Upgraded stylized pipeline with faster inference and improved color accuracy
- Launched referral program for friends who sign up with your link
- Added social share buttons for generated images (Twitter, Pinterest, Reddit)
- Introduced generation history search and filtering by model, date, and prompt keywords
January 15, 2026
Improvement
Negative Prompts & Advanced Controls
- Added negative prompt input field for excluding unwanted elements from generations
- Introduced guidance scale slider for fine-tuning prompt adherence (range 1-20)
- Added seed control for reproducible generation results
- New aspect ratio presets: 16:9, 9:16, 3:2, 2:3, 4:5, 5:4, and custom dimensions
January 5, 2026
Fix
Bug Fixes & Stability
- Fixed issue where free credit counter did not reset at exactly 24 hours for some users
- Resolved Safari rendering issue with generated image previews
- Improved error messages for failed generations with actionable suggestions
- Fixed Stripe webhook handling for plan upgrades taking effect immediately
December 2025
December 18, 2025
v2.0
Video Generation Beta & Pricing Tiers
- Launched video generation with audio beta with ZSky's video engine model for early access users
- Introduced four-tier pricing structure: Free (unlimited, ad-supported), Pro ($19/mo, ad-free + priority GPU), Ultra ($49/mo, 4K + all models), Max ($99/mo, dedicated GPU + white-label)
- Added user dashboard with generation history, credit tracking, and account management
- Integrated Stripe for secure payment processing with support for all major credit cards
December 8, 2025
Feature
Gallery & Community Features
- Launched public gallery with opt-in sharing for community inspiration
- Added prompt visibility toggle — choose whether to share your prompt with gallery images
- Introduced "remix" feature to use gallery images as starting points for your own generations
- Added image metadata display showing model, resolution, and generation parameters
December 1, 2025
Improvement
photorealistic Model Optimization
- Optimized photorealistic inference pipeline for 25% faster generation times
- Improved photorealistic text rendering accuracy for prompts containing words and phrases
- Added support for 1024x1024 and higher resolutions on AI engine
- Reduced GPU memory footprint allowing more concurrent generations
November 2025
November 15, 2025
v1.0 — Public Beta
ZSky AI Public Beta Launch
- Opened ZSky AI to the public with advanced AI image generation
- Free tier: unlimited free generation, ad-supported, no credit card required, 1080p videos with audio
- Web-based interface with model selection, resolution controls, and prompt input
- Google account and email sign-up options for saving generation history
- Published Privacy Policy, Terms of Service, and Content Policy
November 8, 2025
Feature
stylized Integration & Multi-Model Support
- Added stylized as second model option alongside photorealistic
- Implemented model switching in the generation interface — change models without losing your prompt
- stylized offers broader artistic flexibility for stylized, illustrated, and abstract work
- Both models run on dedicated GPU hardware with no shared cloud dependencies
November 1, 2025
Improvement
Infrastructure Hardening
- Deployed automated GPU health monitoring and failover across the cluster
- Added CDN layer for faster image delivery worldwide
- Implemented rate limiting and abuse prevention without impacting legitimate free tier users
- Set up automated backups for user account data and generation history
October 2025
October 20, 2025
Private Alpha
Private Alpha & Core Infrastructure
- Launched private alpha with AI engine on initial GPU cluster hardware
- Built core generation pipeline: prompt processing, model inference, image delivery
- Implemented user authentication system with email and Google OAuth
- Established GPU cluster with dedicated NVIDIA hardware — no cloud rental dependencies
October 5, 2025
Alpha
Foundation & Architecture
- Designed and built the ZSky AI web application architecture
- Set up dedicated GPU inference servers with automatic model loading and management
- Created the generation queue system with priority handling for paid tier users
- Deployed initial zsky.ai landing page and documentation
- Founded by Cemhan Biricik with a mission to make AI creativity accessible to everyone
Never Miss an Update
Subscribe to get notified when we ship new models, features, and improvements. Unsubscribe anytime.
Try the Latest Features
Experience ZSky's video engine video generation with audio, next-gen AI images, and more. Free tier available.
Start Creating Free →