Product Updates
A user-readable update history for new creation tools, public discovery, safety improvements, performance work, and product guidance.
Latest update
Aug 25, 2026
96
Entries
5
Areas
Generation failures now keep internal service details out of the interface while giving creators a consistent, actionable retry message.
Opening the homepage and moving through shared navigation now behaves more consistently across full-page loads and in-app transitions.
Creators on smaller screens get steadier generation controls, history actions, and media presentation throughout the creation flow.
PerchanceAI now exposes MiniMax H3 generation options with matching duration, reference, and pricing capabilities in the model selector.
Creators can choose saved assets or inspect new image uploads directly from the generation composer before starting a task.
Interactive creative tools now share the main generation workspace and expose clearer presets for starting common workflows.
The AI Photo Editor is easier to find and now follows the same generation workspace pattern as other image-to-image tools.
Local image utilities now present a simpler control surface and a more direct explanation of how uploaded media is handled.
Creators can download an image directly while reviewing it in the full media preview, without returning to another page.
Kling V3 can now use motion-control inputs in supported video workflows, with validation aligned to the selected model.
Product-backed image tools once again run through the supported generation path with clearer readiness and model compatibility checks.
Creators can guide video movement from a dedicated workspace with reference uploads, motion settings, previews, and generation results.
Changing languages now reaches the requested localized page more reliably, even when browsers or edge caches retain earlier redirect state.
The image-to-video route now uses the same responsive creation shell as the main generator while keeping its public URL unchanged.
Image and video creation now share a more consistent composer, output settings, prompt assistance, history, and example browsing experience.
Creators can now select multiple media nodes on the canvas board and batch export them into a single organized ZIP archive.
Real-time prompt enhancement suggestions load significantly faster with expanded style presets and negative prompt recommendations.
High-resolution upscaling now restores subtle textures and facial details up to 4K clarity without losing composition.
Video creators can now specify camera pan, zoom, and tilt directions for image-to-video transformations.
Organize creative inspirations using color-coded sticky notes and rich text tags directly on the infinite canvas.
Anonymous preview credits and generation history automatically sync into newly created accounts upon sign-in.
Creators can compare model speeds, artistic strengths, credit costs, and resolution limits side-by-side.
Navigating, zooming, and selecting nodes on touchscreens and mobile devices is noticeably smoother.
Quickly locate saved creations by searching prompt keywords, date ranges, or aspect ratio tags.
Strengthened safety guardrails protect public browsing while offering clearer guidance when prompts touch policy boundaries.
Generate video clips formatted directly for social media stories, vertical video feeds, or YouTube widescreen.
Revert node repositioning, content updates, or parameter changes with comprehensive undo and redo support.
Explore public community creations and instantly import generation parameters, seeds, and prompts into your generator.
Media previews and generated output thumbnails load up to 60% faster across all global regions.
Creators can now browse a much broader set of visual workflows and practical guides before choosing the right PerchanceAI creation path.
Canvas boards are clearer in light theme, preserve media more reliably, and support a more direct first-frame video workflow.
Creators can assemble clips from mixed video sources, preview the combined result, and continue working without leaving the board.
Extracting clips and arranging connected media now takes fewer steps, with clearer controls across different canvas sizes.
The newer canvas workspace handles more list, text, connection, frame, and display tasks with smoother large-board behavior.
Creators can use a more responsive canvas foundation for selecting, connecting, moving, and reviewing nodes.
Canvas video workflows can combine clips with audio and pass clearer frame roles into supported generation models.
Creators can combine multiple videos more reliably and see clearer progress while prompts are being optimized.
Working with batches of images and mixed media in Canvas is faster, clearer, and more dependable on narrow nodes.
Creators can clean visible watermarks from generated videos through Image Tools, using the new Omni workflow for videos they have the right to process.
Canvas boards feel clearer when browsing saved spaces, resizing images, and choosing video start or end frames as references.
Longer canvas work is less likely to lose context when generating media, reconnecting nodes, or returning after an offline recovery.
The canvas received a broad performance and interaction pass so larger boards can remain easier to move through, edit, and review.
Creators can duplicate canvas projects and work with media nodes through clearer title, port, group, and save behavior.
Building from existing nodes is more direct, and canvas editing feels clearer in light theme and restricted browser environments.
We are preparing a larger visual workspace where creators can arrange references, generated images, notes, and follow-up edits without losing the thread of a project.
Canvas work is moving toward clearer side controls, so creators can add, review, and adjust content without breaking their flow.
Annual members should be able to understand that plan credits are intended for monthly creative work instead of feeling like a single unclear annual grant.
Generated images and uploaded references are being connected more clearly to upcoming canvas work, so useful assets can move into longer projects with less repetition.
We are planning starter canvas layouts for common creative jobs so users can begin from product shots, posters, references, or video source frames faster.
The infinite canvas is being shaped around longer sessions and larger boards, with performance work focused on keeping movement and review comfortable.
Creators should be able to swap an image and continue a direction without rebuilding the whole setup, especially inside longer visual projects.
We are reviewing longer creative sessions so the upcoming canvas can handle heavier work without making the workspace feel stuck or unreliable.
Moving around a larger board should feel predictable, with clearer selection, panning, and return-to-edit behavior for detailed creative work.
The upcoming canvas is being planned around real creative tasks: generating images, comparing references, preparing video frames, and finishing assets.
Public browsing now includes fresher topic pages while private app areas stay more clearly separated from visitor-facing discovery.
The home hero now speaks differently to new visitors and signed-in creators, making the next step clearer for each user.
New visitors can see the sign-up credit offer sooner, helping them understand what they can try after creating an account.
Video creation now has a clearer place in navigation, and shared footer links behave more reliably when visitors browse from partner or reference pages.
Explore, prompt, and tag browsing received a quieter copy refresh so visitors can understand public examples faster.
Prompt and inspiration pages were refined so users can browse ideas with clearer topics, cleaner labels, and more practical starting points.
When a generation fails and you try again, PerchanceAI now keeps the model and options from the original attempt so the retry feels predictable.
Monthly and annual Plus plans now show credit timing more clearly, and the pricing page does a better job identifying your current plan.
Visitors are more likely to land in the right language, while public guide, tag, and image-to-video pages provide clearer browsing paths.
Creators can now redeem promo codes from pricing, understand code problems more clearly, and see member status more visibly in the workspace.
Image Tools and public guide pages now give clearer examples and guidance. Personal media browsing and uploads are also more responsive.
The video page now gives creators a clearer starting point. Public pages and prompt galleries show cleaner titles, safer descriptions, and more consistent visible counts.
Image Tools are easier to discover, mobile result actions are easier to continue from, and video tasks wait more reliably until a real result is ready.
Creating on phones now feels more stable: the workspace, generate bar, video covers, and creation links are easier to use on small screens.
The home page handles uploaded images more clearly, sign-in brings users back to the right place, and personal generated works can be removed when they are no longer needed.
Account creation is more reliable, automated sign-ups are checked more carefully, and new users get a clearer welcome-credit experience.
Image and video creation now start from a more consistent workspace. Reused prompts travel more smoothly, and unusually long waits are handled more fairly for credits.
Public browsing now shows cleaner work details and more useful idea pages, helping visitors explore styles, use cases, and prompts with less noise.
Sign-in now keeps you closer to the creation flow. The generator shows choices that fit the selected model, and switching languages feels lighter.
Credit previews now better match the selected model and size. Prompts entered before sign-in are preserved, and the home page includes a World Cup themed creation entry.
The home experience is more focused on starting a creation. Public gallery previews show cleaner information, while video previews and sidebar navigation feel smoother.
Credit and pricing previews are more consistent, public tag pages are cleaner, and shared footer links are safer to browse.
Quality and resolution choices now adjust to the selected model. Credit pricing can stay clearer as model options change, and PerchanceAI now has a public About page.
The pricing page is easier to compare, Explore videos feel more alive, and Image Tools previews plus sign-in flow are clearer.
The home page now centers on the prompt box, Image Tools, example results, pricing, and helpful footer links so new visitors can start faster.
Creation pages now communicate task progress and media loading states more clearly, reducing uncertainty while results are being prepared.
Image Tools now show cleaner before-and-after examples and clearer guidance across supported languages.
Localized visitors can reach the correct Image Tools pages more reliably, including tools such as AI Image Extender.
Anonymous users get a clear sign-in prompt as soon as they try to use protected Image Tools.
The main Generate history is easier to scan because background removal, image extension, and other Image Tools jobs stay in their own tool context.
The image extension tool now uses the clearer AI Image Extender name, while older AI Uncrop links still open the updated tool safely.
Rapidly switching between image and video tabs in Explore now keeps the selected view stable and predictable.
The canvas layout and edit flow are clearer, with fewer extra steps before starting an erase or object removal task.
Three standalone utilities now run in the browser with no server upload and no credit usage.
The sidebar is easier to scan, with clearer sections for discovery, image creation, video creation, and personal space.
Users can now perform precise edits such as background removal, upscaling, object erasing, and outpainting directly on generated or uploaded images.
Users can find product guidance and track what changed without contacting support first.
Public pages are safer because private review labels stay out of Explore and public media details.
Generated assets can be organized with clearer public tags while private review information stays separate.
Users can browse public work with clearer previews and popular views by day, week, or month.
Signed-in users can follow creators and build a feed from public work they care about.
Users can create short videos alongside image generations when video models are enabled.
PerchanceAI combines AI image generation, local image utilities, product photo workflows, and image-to-video tools in one creative studio.
Introducing PerchanceAIThe Help Center explains how to use current features and what to send support when something does not look right.
Open Help Center