What's New

The subscribe page says what every plan includes. Subscribers can buy more credits right from it.

The community gallery on the home page shows each tool's real name.

Settings is redesigned. See your plan and credit balance in one place, and turn email types on or off right there.

Settings shows when a Civitai key is saved, and you can remove it.

The Models menu has a new Edit & enhance column, so editing, upscaling and camera-angle tools are one click away.

The Effects menu lists the effects people use most, starting with AI Kiss and AI Hug.

Academy is now a menu of its own: lessons and tutorials, plus recipes, prompts, presets and the community gallery for ideas.

Model lists are ordered by how many people use each model, so the menu shows what most creators pick.

The Sensei page now says which AI does what, and that a take that misses is made again or flagged for you to decide.

Kata edits show their rough credit cost on every card, and the season's edit comes first.

Search Video Effects by name. Type a word above the grid to find an effect fast.

The pricing page says up front which four tools are unlimited, and counts image and video models the same way.

8 Video Effects are back, including Back Hug, Mermaid Tail and Pet Pilot.

Video Effects run on a newer engine. Your photo stays exactly as you took it at the start of the clip, and faces hold up better.

Two-person effects like AI Hug keep both people looking like themselves.

MiniMax H3 Turbo makes a video from words alone again. A photo is now optional.

How It Was Made pages are easier to read. Each one shows its date and rough cost up front, with plain settings and short notes.

Sensei Recipes are easier to browse. Filter by ads, holidays, real estate and more, and see the rough credit cost next to Run.

Go Viral is open. Make anything with Pixel Dojo, post it, and the 10 most-viewed posts share $2,450 in cash.

Enter from your Creator dashboard and pick the Library item your post was made from. Every approved post still earns $1.50 per 1,000 views.

Film Studio: add a scene anywhere in your storyboard. Say what happens and only that scene is planned. Every scene and clip you already made stays as it is.

Short films: give a prop or an extra place your own picture before you approve the price, just like the main location.

P-Video Upscale makes any clip sharper at 720p, 1080p, 2K or 4K, and keeps the sound. Find it under Make it sharper.

Pick 60 fps for smoother motion. Upscales start at 0.3 credits a second.

Every reference image now shows in the composer, even past ten. No more hidden pictures behind a +N tile.

Presets now run on our newest models: FLUX 3, Nano Banana 2.1, Ideogram 4.5, Seedream 5, Kling 4.0 and Veo 3.1 Fast. Each one has a fresh example.

Drag the corner of the prompt box to resize it. Make it smaller to see your results, or bigger to read a long prompt.

Upload reference images up to 50MB, so full 4K files go straight in.

Seedream presets now open in the shape they were made for, like 16:9 for a street scene.

On laptop screens, the prompt box and Generate button now show without scrolling on image and video pages.

Recent moved into the top menu, so it no longer covers the prompt box's buttons.

Dojo Academy remembers where you left off. Every track and the start-to-finish path now say Continue and open your next lesson.

The community gallery is cleaner. Videos name the model that made them, and empty settings and zero counts are gone.

Trending now covers the whole week and only lists posts people reacted to. Leaderboards say what each number means.

Creator profiles lead with the creator's name and fit a phone screen, so their work shows right away.

The model library now compares the newest models with one shared prompt, including Seedance 2.5, Kling 4.0, WAN 3.0, FLUX 3 and Seedream 5.

Prompting guides for the newest models come first, with simple filters. Guides for older versions are tucked below.

Use Cases topics now hold the pages that match them, with a picture on every topic and cleaner page titles.

The contests page says when no contest is open, and points you to the Creator Program to get paid for views in the meantime.

The affiliate promo kit now links and names the newest models, with a fresh video you can repost.

Brand kits now live in your Library under Brands. Save your logo, colors, fonts and voice once, keep as many brands as you need, and pick a default.

Your brand fills in on its own in Brand Overlay, video captions and Faceless Studio. Your logo is one tap away in Kata and in every image picker.

Film Studio captions now use your brand color, and you can add your brand end card after any finished film.

Key pages now read natively in French, Korean, Indonesian and Turkish too, from the home page and pricing to effects, tools and our comparisons.

The LoRA trainer page and nearly 100 of our most-read guides are now in all 13 languages.

Character Studio is simpler: your character's photos sit in one panel, Scenes, Sheet and Dance are tabs, and the composer stays at the bottom once you have shots.

Picking characters, places and props is now a searchable grid with big pictures, the same everywhere: Character Studio, Film Studio, Sensei, Kata, Character Swap, ads and image uploads.

Character Studio works much better on phones, with a roomier prompt box and everything important on the first screen.

Saved prompts have one clear button with a menu: save this prompt, or use a saved one.

Your prompt and attached cast no longer carry over when you switch characters, and removing a photo or deleting a character now asks first and works in every browser.

The ad studio is simpler: pick a product, pick a style from big examples, and press Generate. Fine-tuning like the opening, scene, quality and length now lives under Customize.

Film Studio is calmer: one Options button instead of three rows of choices, tidy film cards with clear progress, fewer notices on each film, and plain words throughout.

Kata has a clearer start: pick a ready-made edit or use your own video, then follow four short steps. The Plan button is always in reach, and every ready-made edit shows its picture.

Character Swap is simpler: one tap picks who stars, it tells you what's still missing before you can swap, and the choice is just Standard or High quality.

Video Studio is easier to follow. Pick Make, Change, Move or Extend first, then add what it needs. Try this on an example loads its video, ready to generate.

Faceless Studio shows an episode's full cost before you start, and you can hear each narrator first. Finished episodes open right on the video.

Music videos are clearer: the price is worked out from your song's real length, long songs get a heads-up, each look says what it looks like, and every music video is named after its song.

Film Studio now starts by asking what you're making: an idea, a micro-drama, your script or a music video. Each one shows a real example, and the composer only asks what that kind needs.

Prompt Studio says what it does, folds away one-line drafts, warns before an old draft is deleted, and asks before deleting a version. Each part of a prompt can be edited on its own.

Saying yes to a Sensei quote now approves it. Micro-dramas and music videos always show an Approve button for their real price, so they no longer stop halfway to ask again.

Canvas is tidier: your image is no longer hidden behind the composer, WAN 2.7 Image Edit and Kling Video show proper size and quality choices, and the model list is simpler to scan.

Tool pages open faster: the example pictures load in a fraction of the size. The Generate button shows credits with the ⚡ icon, and Seedream 5 shows what each size and version costs.

The Pricing page now shows what Kling 4.0 and Nano Banana 2.1 cost, and says which models are unlimited on every plan.

The Pricing page in all 12 other languages now shows the same prices as the English one.

Our pages in other languages now feature Kling 4.0 alongside the other top video models, and count all 310+ video effects.

Every preset in the Preset Library now opens its tool with the prompt and settings filled in. Before, the FLUX and Nano Banana presets opened an empty page.

Photo Effects put the holiday in season first, and the look picker now just says Recommended or Other look, with a note when the other one costs less.

AI Tools are easier to browse: jump straight to a group, video tools play their examples, and each tool tells you what it still needs before you can press Generate.

Your Library can show only images or only videos, folders are searchable, and you can move or delete an item right from its view.

Kling 4.0 no longer turns down requests up front that it can make. Your prompt goes straight to the model.

When Seedance can't fetch one of your references in time, it now keeps trying for longer before giving up, so far fewer runs fail on a slow connection.

An old generation no longer pops up a "Generation timed out" notice again and again as you move around the site.

A character you build in Character Studio now opens ready to shoot, with one-click options to use them in Sensei, Film Studio, an ad, or a dance and effects video.

Character Studio has a new "Make something with" button that opens Sensei with your character ready, plus sharper saved shots and clearer tips.

Every tool now has a Feedback on this tool link under the prompt box. Tell us about a missing setting, something confusing or a bug, and it goes straight to the team.

Ask Sensei for a screen tutorial and the finished video shows up in the same chat. No need to ask whether it is done.

Your recent results keep their settings, like length, quality and voice, when you work in more than one tab or make a lot in a day.

Tall videos open full size with Save and Download in view.

Marketing Studio frees the Generate button as soon as your first image starts, so one slow image no longer holds up the next ad.

Tap thumbs down on a result and you get a fix right away: one tap on Enhance rewrites your prompt and puts it back in the box, ready to run again.

Clearer tips for the most common problems, written from what you told us.

On video tools, a short prompt now gets a gentle nudge to add who, what, where and the camera move. Short prompts are the ones that most often get ignored.

Run out of credits partway through a film and Film Studio shows what is left to make and what it costs. Add credits and the film picks up right where it stopped.

The latest frame and clip of every shot is kept for you, so a film you leave overnight is all there when you come back.

Simple mode films no longer stop over small slips in the written plan, and an idea the writer can't make now says so plainly, with your idea ready to change.

Sensei keeps every photo you attached in mind for the rest of the chat, so "use all four photos" uses all four.

A song or clip longer than a video model takes as a reference is trimmed for you instead of stopping the shot, and Sensei tells you which part it used.

Plans lead with what the project costs. Clips that start from a frame you already have no longer add the price of a new one.

Belts are within reach: 5 beginner lessons earn your White belt, then 6 for Green and 3 for Black. Every lesson shows how close you are.

When a lesson ends, the next one is ready and starts on its own, and the Academy picks up where you left off on the hub and your dashboard.

Learn how on each tool page now opens the lesson about that tool.

A new lesson, Make your first AI image: describe four things, pick the shape, make two takes and keep the best.

Connecting ChatGPT to pixeldojo.ai/mcp now finishes: approve once and your Pixel Dojo tools show up in ChatGPT, and the connection stays signed in.

Add to Claude: one click opens Claude with PixelDojo ready to connect. Make images, video and ads right in your chat.

Add to ChatGPT: one click copies the PixelDojo connector and opens ChatGPT, so you just paste and connect.

Studios, Effects, Tools and Models now open the same kind of menu, so everything sits in the same place wherever you look.

The Models menu leads with the top image and video models, the Unlimited ones first, plus a quick way to browse by what you want to make.

On phones and tablets the menu has the same sections as on desktop, in the same order, with the same shortcuts inside.

Sizzle reels now play at the frame rate your clips were made in, so motion is smooth: no stutter, no repeated frames, and footage plays as filmed while stills get the slow push.

A shot that comes back in a sizzle reel picks up where it left off instead of jumping back, and no cut lands inside a clip's own edit, so there are no split-second flashes.

Sensei makes storyboard sizzle reels from 15 to 60 seconds, each shot with its own headline. Ask for before and after pairs to show off a product.

A song or score the music model turns down now fails right away and is refunded, instead of spinning until it times out, and a sizzle reel simply makes a new one.

Video effects and our fast video model now render about three times faster, so a 5-second clip is usually ready in well under a minute.

We re-tested all 310 video effects. 21 now follow their description more closely: no stray lettering on clothes and signs, the right number of people in group shots, and camera moves that really move.

Video effect soundtracks are instrumental again: no more garbled singing over your clip.

AI Muscle now runs on our own fast video model: 7.5 credits instead of 30, and it works on the free tier and Relax too.

Enhance now knows when you attached reference images: it sharpens your edit instead of rewriting it into a new scene, and points at your images rather than describing them again.

Enhance no longer adds camera, lens, blur or lighting details you never asked for on Seedream and other image models.

Sensei now opens like our other studios: your prompt front and center, with skills you can scroll through right under it and every recipe shown with a real example.

The recipe picker shows a picture for every recipe, including the ones you saved.

Every image and video result has a thumbs down. Tap it, pick what went wrong, and add a note if you like.

Each reason comes with a tip for fixing it, a shortcut to edit your prompt, and that model's prompting guide.

Enhance learns from what you keep and what you mark down, so its rewrites get better over time.

Refine now works on long enhanced prompts. Before, a prompt past about 4,000 characters showed Working… and then changed nothing.

On phones, when a refine can't make a new version, the reason now scrolls into view instead of hiding under the prompt.

MiniMax H3 Turbo now builds every video from your pictures: start from a photo, or add up to 9 reference images, clips or sounds and call them out in your prompt.

One simple price: 7.5 credits for 5 seconds at 720p, 3.75 at 480p, and still free on Relax.

Video effects run on the upgraded model, with sharper faces on favorites like AI Hug, Old Photo Animation and Santa Suit. Every effect is now 7.5 credits.

In Video Studio, Change your clips with MiniMax H3 Turbo: swap an outfit, the background, the time of day or the weather, and keep every move and camera angle.

Run out of credits and the message now points you to Relax, where Qwen Image 2.1, FLUX 3 Klein, MiniMax H3 Turbo and the effects run with no credits.

Right after you subscribe, you see everything that is unlimited on your plan, one click from each.

Model pages and comparison pages put the Create box near the top on phones, and comparison tables fit the screen.

On phones, Pricing opens on the plans headline and the Subscribe button.

Signing up no longer jumps as it loads, with bigger buttons and fields that don't zoom in on iPhone.

Film Studio shows your full film titles on phones.

The site footer folds into sections on phones, so pages end where they should.