A complete walkthrough of Studio, Header Studio, Content Studio, and Lab. Covers credits, prompt writing, and the small tricks that save hours.
Quick Start
Pick a tool and create your first design in under a minute. ReelyArt is built so you can ship social-ready visuals without any design experience.
1Sign up with Google or X. 10 free credits land in your account instantly.
2Choose your tool: Studio for images, Header Studio for banners, Content for GIFs/memes/stickers, Lab for NFT scenes.
3Type a prompt, optionally attach reference images, and press generate. Edit the result or regenerate as needed.
New to ReelyArt? Start with Studio. Upload one image, run a curated template or write your own prompt, and you have a finished visual in a few seconds.
StudioImage-to-image at /studio. Pick a template, upload your photo, get a stylised result. Or stay in Custom mode and write your own prompt with up to 8 reference images, 10 on FLUX 3.
Header StudioBanner and header generation at /studio/header for X, Discord, and websites. Pure text-to-image at preset dimensions.
Content StudioShort-form assets at /studio/content: animated GIFs, reaction memes, and transparent sticker emotes.
LabNFT-aware scene generator. Pulls live OpenSea data and renders cinematic compositions around your collections.
Credits & Plans
Credits are the only currency on ReelyArt. Every AI generation deducts credits up front and refunds them automatically if the generation fails.
Content: GIF, clips longer than 4s (5-8s, paid plans)
price x seconds / 4
Content: Meme edit (chat-style revision)
2
Lab: NFT scene (1K)
3
Lab: NFT scene (2K)
4
Lab: NFT scene (4K)
6
Edit an existing image (Studio / Header / Lab)
Same as one image at the selected quality
Plans
Plan
Credits / month
Price
Free
10 (signup bonus, with Google or X)
$0
Starter
100
$9/mo · $7/mo yearly
Plus
250
$19/mo · $15/mo yearly
Pro
550
$39/mo · $32/mo yearly
Max
1,450
$99/mo · $80/mo yearly
Subscription creditsReset at the start of each billing cycle. Unused credits do not carry over.
Credit packsOne-time purchases that never expire. Always spent before subscription credits.
Changing your plan
Upgrade: monthlyTakes effect immediately. You are charged a one-time invoice for the exact difference: the new plan's full monthly price minus the proportional value of your remaining unused subscription credits. Your new credit allocation is granted right away.
Upgrade: monthly to annualSwitching to an annual plan is always available and always treated as an upgrade, regardless of tier. You are charged the full annual price minus a discount based on your remaining monthly credits. Your new monthly credit allocation starts immediately.
Upgrade: annual to annualTakes effect immediately. The remaining months on your current annual plan are credited as a discount toward the new annual plan charge. Your new monthly credit allocation starts immediately.
Downgrade: monthlyScheduled for your next billing period. Your current plan, features, and credits remain fully active until the cycle ends, then the lower plan takes effect.
Downgrade: annualNot available mid-year. Annual subscriptions cannot be switched to a monthly plan or a lower annual tier before the renewal date. Select a new plan to apply at your next annual renewal.
CancelYou can cancel at any time from your Account page. Cancellation takes effect at the end of your current billing period, access, features, and credits remain fully active until then. Annual subscriptions stop renewing at the end of the paid 12-month period. No partial refunds are issued for unused time.
If you run low mid-project, the Pricing page offers one-time credit packs starting at a few dollars. Pack credits never expire.
Studio
Studio is the image-to-image workspace at /studio. Your gallery fills the page and the composer sits at the bottom of it: write your own prompt with up to eight reference images (ten on FLUX 3), or switch to Template and run a curated look over one photo.
Three Studio surfaces
StudioImage-to-image transformations. Pick a template, upload your photo, get a stylised version back. Default landing at /studio./studio
Header StudioGenerate ready-to-use banner art for X, Discord, and websites. Pure text-to-image at preset sizes./studio/header
Content StudioGIFs, memes, and transparent sticker emotes. Short-form social assets./studio/content
How a Studio generation works
1Choose a mode in the composer at the bottom of the page: Custom for your own prompt, Template for a curated look.
2Add your images: drag them anywhere on the page, paste from the clipboard, or attach one you uploaded before.
3In Template mode, open the Templates tab to pick the look. Search, filter by category, or sort by trending.
4Pick an aspect ratio, a quality and a variation count (1 to 4). The first two lists are the model’s, not the page’s. See the table below for what each one offers.
5Generate. Credits are deducted up front and refunded if the generation fails.
6Results land in History, the gallery filling the page behind the composer.
The two modes
CustomYour own prompt with no template on top of it, plus up to 8 reference images (10 on FLUX 3), or none at all for text-to-image. You pick the model here: Nano Banana 2, 2.1 or Pro, GPT Image 2.5 Flare or Sunburst, GPT Image 2 or 1.5, Seedream 5 Pro, 5 Lite or 4.5, Grok Image 2, FLUX 3, Ideogram 4.5 and Krea 3. Each one renders its own set of sizes and shapes, so the quality and ratio dropdowns follow the model you picked. On the free plan you can run twelve of them at their opening tier: Nano Banana 2 and Nano Banana 2.1 up to 2K, GPT Image 2 at 1K, all of GPT Image 1.5, GPT Image 2.5 Flare and Sunburst at 1K, Seedream 5 Lite and Seedream 4.5 at 2K, Grok Image 2 at 1K, FLUX 3 at 768 and 1K, Ideogram 4.5 at Low, and Krea 3 at 1K Instant. Nano Banana Pro, Seedream 5 Pro and the higher tiers of the rest are on paid plans.
TemplateA curated look, written and tested for one job. Add a single image and send; the template carries the instructions and the model it was built for.
Templates that take wordsA few templates put your text in the image: a name, a greeting. Those show a text field in the composer; the rest do not, because anything typed would reach the model as extra instructions.
The fourteen models, side by side
Every model renders its own set of resolutions and shapes, at its own price, so the quality and ratio dropdowns change with the model you pick. Nothing is hidden behind a generic tier name: a resolution listed here is the resolution the file comes back at.
Pick 1 to 4 variations per generation; the total is the tier price × the number of variations. Switching to a model that lacks the tier or shape you had selected snaps to the nearest one it does have rather than refusing the request, and keeps your original choice, so switching back restores it.
Nano Banana 2.1: the update to Nano Banana 2
Nano Banana 2.1 is Google's update to Nano Banana 2. It follows long, layered prompts more closely, sets text inside the picture more cleanly, and keeps a character consistent across follow-up edits. It renders the same three sizes as Nano Banana 2 for fewer credits at 2K and 4K, and a free plan runs it up to 2K.
Tier
Output we measured
Credits
Plan
1K
1024 × 1024 at 1:1
2
Free
2K
2752 × 1536 at 16:9
3
Free
4K
5504 × 3072 at 16:9
4
Paid plans
Write the whole briefName every object, where it sits and what colour it is. Nano Banana 2.1 is built to keep the details of a long prompt, so a precise one pays off.
Quote the wordsPut any text that should appear in the image in quotes, such as a menu board that reads "RAMEN 12".
ReferencesUp to 8 reference images on paid plans and 4 on the free plan. Say in the prompt which image plays which part.
TimingA 1K image takes about 15 seconds, 2K about 20 and 4K about 30.
FLUX 3: five sizes and precise edits
FLUX 3 is the newest image model from Black Forest Labs, built for editing as much as for generating. It renders five real output sizes, takes up to 10 reference images, and on an edit it changes what you name while usually leaving the rest of the picture as it was. The composer always shows the price you will pay before you send.
Tier
Output we measured
Credits
Plan
768
768 × 768 at 1:1
1
Free
1K
1360 × 768 at 16:9
2
Free
1.5K
1536 × 1536 at 1:1
3
Paid plans
2K
2736 × 1536 at 16:9
4
Paid plans
4K
5456 × 3072 at 16:9
20
Paid plans
1Text to image: choose FLUX 3 in the model menu, write the prompt, pick a ratio and a tier, and send.
2With references: attach up to 10 images and say in the prompt which image plays which part, for example "the jacket from image 2 on the person in image 1".
3Edit: open a FLUX 3 result, press Edit and describe only the change. Keep the ratio the image already has and the change happens in place; pick another ratio and the scene is recomposed to the new shape.
Describe the change, not the sceneFor an edit, "make the mug matte black" works better than describing the whole picture again.
Quote the wordsPut any text that should appear in the image in quotes, such as a sign that reads "OPEN".
Be specificName the style, the light and the framing. FLUX 3 follows detail closely, so a precise prompt pays off.
Draft small, finish largeFind the idea at 768 or 1K, then run the one you like at 2K or 4K. A 1K image takes about 20 to 30 seconds; 4K about a minute and a half.
Ideogram 4.5: type and design
Ideogram 4.5 is the model to reach for when the words in the picture matter: posters, labels, menus, covers and social graphics, in many languages. It also edits: change a label, recolor a product or swap an object and the rest of the picture stays put. Up to 5 reference images.
Tier
Credits
With an image
Plan
Low
2
2
Free
Medium
3
3
Paid plans
High
5
8
Paid plans
Tiers are effort, not sizeFrom words every tier renders at 2K (2048 × 2048 square, 2560 × 1440 wide). Medium and High spend more effort on fine detail and small type; Low is already clean for most posters. High costs more with an image attached, because Ideogram bills it that way.
Shape from words, or from your imageFrom words alone you pick the ratio. Attach an image, or edit one, and the result takes that image's shape and size, so the ratio menu shows Auto and a small upload gives a small result.
Quote the wordsPut the exact text in quotes and say where it sits: the title at the top, a small line at the bottom.
Krea 3: fast, instruction-led edits
Krea 3 is Krea's image editing model. You say what should change and what should stay the same, and it makes the edit; it also generates from words alone. Each tier pairs a size with a thinking mode: Instant goes straight to the render, High plans the edit first, for complex changes or several references combined. Up to 4 reference images on paid plans, 1 on the free plan.
Tier
Output we measured
Credits
With images
Plan
1K Instant
1024 × 1024 at 1:1
2
3
Free
1K High
1024 × 1024 at 1:1
3
4
Paid plans
2K Instant
2048 × 2048 at 1:1
3
4
Paid plans
2K High
2048 × 2048 at 1:1
4
5
Paid plans
Lead with the changeSay what changes first, then close with one sentence that keeps the rest, such as "keeping everything else the same". Listing everything that should stay is not needed.
Name each imageWith more than one reference, refer to each by what it shows: "put the sneaker from the second image on the person in the first".
ShapePick a ratio and the result is framed to it. Edit an image and keep its own ratio, and the picture keeps its shape.
When to use HighStart on Instant. Move to High when an edit has several steps, combines references, or Instant missed part of the request.
Custom prompt
Skip the custom promptFor standard templates, the empty prompt gives you the cleanest, most predictable result.
Add a few keywordsTo override colors, mood, or specific elements: "darker tones", "smiling", "winter setting", "neon background".
Required in Custom modeCustom mode has no preset prompt, so you describe what you want from scratch. The sparkle button expands a short one before you spend a credit.
Tips for better results
Input quality mattersStudio works best with clean, well-lit source images. Avoid heavy JPEG compression, watermarks, and busy backgrounds.
Favorite templatesTap the heart on any template to save it to the Favorites tab for one-click access next time.
Use variations, not rerunsIf the first result is close but not perfect, bump variations to 2 or 4 instead of running the same prompt twice.
History is the pageEvery generation fills the gallery behind the composer, and the slider top-right sets how many fit per row. Each card carries the template it came from, click that badge to load it again.
Edit any result
Every finished image has an Edit action, the pencil button on the card (inside the three-dot menu on mobile) or the Edit button in the fullscreen view. It opens a chat bar under the image: describe the change in plain words ("make the background night", "add falling snow") and send. Your instruction goes straight to the model with no template prompt, and the revised image lands in your gallery as a new card. The original is never overwritten. The edit runs on the model that made the image, so it costs the same as a single generation at the selected tier and offers that model’s tiers and shapes. Grok Image 2 is the exception: its input schema ignores the aspect ratio once an image is supplied, so an edit there keeps the original’s shape and the bar offers no ratio control; Ideogram 4.5 works the same way. FLUX 3 is the model built for this: name one change and it usually leaves everything else where it was.
New to Studio? Pick a template marked NEW first. These are the most recent additions and usually showcase the strongest results.
Header Studio
Header Studio at /studio/header generates banner art for X, Discord, and websites. No layers to arrange. Pure image at the dimensions you need, ready to download.
X Header1500 × 500 at 3:1, for X / Twitter profile banners.
Discord Banner960 × 480 at 2:1, for Discord server banners.
Website Header1920 × 600 at 16:5, for landing pages and hero sections.
Wide Header1920 × 480 at 4:1, for ultrawide social headers and section dividers.
Quality and credit cost
Quality
Resolution
Credits per image
1K
Standard
3
1K
Fast model (opt-in, faster & cheaper)
2
2K
High detail
4
4K
Maximum detail and sharpness
6
You can request 1, 2, 3, or 4 variations per prompt. Total cost equals quality credits × variation count. The Fast model is an optional 1K-only toggle in the chat bar for quicker, lower-cost banners.
Style presets
Header Studio ships with a curated style dropdown including comic, anime, cinematic, minimal, photographic, and more. One special mode worth knowing:
Text BannerRenders your nickname or handle as the primary visual element. Best for personal X profile headers and Discord identity banners.
Reference images and crop tool
Reference imageDrag, paste, or click to attach a single reference. Header Studio uses it as a strong anchor for composition, colors, and mood.
Crop toolAfter generation, click any header to fine-tune the crop. Shift the focal point horizontally or vertically before downloading.
Edit any banner
Hover a card and use the pencil button (or press Edit in the fullscreen view) to revise a finished banner with a plain-language instruction, "make it night", "add more neon". The revised banner is saved as a new card next to the original. Quality is selectable in the edit bar (the wide banner ratio stays fixed) and an edit costs the same as one banner at that quality.
For X profile headers, leave clear space on the left for the avatar circle. Mention this in your prompt: "leave the left third clean".
The Content Studio at /studio/content is built for short-form social assets. Three modes are available: GIF, Meme, and Sticker, each tuned for one specific job.
Generate an animated GIF from a reference image plus a motion description. Pick a model, a quality tier, a clip length and a ratio (square 1:1, portrait 3:4 or landscape 4:3; every ratio costs the same) before generating. Seedance 2.0 Pro renders 480p (10 credits), 720p (25 credits), or 1080p (50 credits). Seedance 2.0 Mini costs about half, 480p (6 credits) or 720p (12 credits), and tops out at 720p. Seedance 2.5 is the newest model, 480p (12 credits) or 720p (28 credits), and the whole model is reserved for paid plans. On Pro and Mini the tiers above 480p require a paid plan; 480p is available to everyone. Higher tiers render at a higher resolution for sharper detail. A slider sets the clip length between 4 and 8 seconds; the price scales with it, so a 6-second clip costs one and a half times the 4-second price and an 8-second clip costs double, rounded to whole credits. Lengths above 4 seconds are a paid-plan feature, and longer clips make heavier GIFs, so 4 seconds keeps the most detail. Every result also comes as an MP4 with generated audio, while the GIF itself stays silent; the speaker button next to the model dropdown turns that audio off if you want a silent video.
1Upload one reference image. This is required.
2Describe the action or reaction you want, or leave it short and let the AI choose a motion that fits the picture.
3Submit. A 4-second 480p GIF takes about two to three minutes; longer clips and higher tiers take a little more.
The image already tells the model what is in the frame. Spend your prompt budget on the action: "blinking and smiling", "hair blowing left", "slow zoom in". To keep a finished picture exactly as it is, say "animate this image" or name the one part that should move ("only the water moves"). Ask for a walk, a run or a flight and the camera follows the character while the scenery streams past; ask for a walk cycle on a plain background and it stays in place.
Every finished GIF can be exported three ways from the card menu or the preview lightbox:
Standard GIFThe optimized GIF created automatically when generation finishes, tuned for fast loading and easy sharing. Free.
Download MP4The original generated video clip, for when you want the highest-fidelity source file.
HQ GIF, ProA maximum-quality re-encode at 640px (552 × 736 for portrait, 736 × 552 for landscape), 20fps and 256 colors, generated on demand and kept in your gallery as a fourth "HQ GIF" tab. It targets the best possible quality regardless of file size, and higher-resolution source tiers (720p/1080p) make it noticeably crisper. Available on any paid plan with no extra credit cost.
4 credits per variation. Choose 1, 2, 3, or 4 variations before generating, all run in parallel with GPT Image 2.
1Upload a reference character or describe one from scratch.
2Describe the meme situation: "shocked face when the code compiles first try", "confused dog math meme style".
3Select how many variations you want (default: 1). More variations means more options to choose from.
Finished memes can be revised for a flat 2 credits: use the pencil on the card (Edit in the card menu on mobile) or the Edit button in the fullscreen view, describe the change in plain words, and the revised meme lands in your gallery as a new card, the original stays untouched.
Sticker
Pick a model before generating. Both deliver clean transparent PNGs ready for Discord, Telegram, or any chat platform. They just get there differently.
GPT Image 2 (4 credits per variation)Generated on a white background, then the backdrop is automatically removed for a transparent cutout.
GPT Image 1.5 (3 credits per variation)Outputs a native transparent PNG directly, with no separate removal step. Cheaper and a strong fit for emote and sticker art.
1Choose GPT Image 2 or GPT Image 1.5 in the model selector.
2Optionally upload a reference image. Unlike GIF and Meme, references are not required here.
3Describe the emote: "cute dragon holding a GG sign", "pixel hamster celebrating".
4Select how many variations you want (default: 1), all run in parallel.
5Download at 128, 256, 512, or 1024 px.
Stickers can be generated without any reference, but uploading one dramatically improves character consistency across an entire emote set.
Lab is a specialised pipeline for NFT-themed visuals. It connects directly to OpenSea, pulls live data (floor price, volume, sales, ownership), and renders cinematic scenes around your NFTs and collections.
2Type your prompt. Mention collection names, wallet addresses, or the analysis you want.
3Use the + menu to search an OpenSea collection by name and attach it as a chip, or to upload a PFP the system will use as your character in the scene.
4Optionally pick an analysis type pill to lock the mode instead of letting the AI infer it from your prompt.
5Set quality (1K / 2K / 4K), aspect ratio, how many variations to run, and whether prices render in ETH or USD, then generate.
Analysis types
Seven modes are supported. The AI detects the right one from your prompt, or you can pick a type pill in the chat bar to lock it.
AnalyzeDeep dive on one collection: floor price, volume, owners, supply, and more."analyze Azuki", "deep dive on X"
7D AnalyzeSeven-day sales report for one collection: total sales, average sale price, and 7d volume as headline stats."weekly report", "last 7 days", "7 day recap"
CompareSide-by-side of 2 to 5 collections across one or more metrics."compare X vs Y", multiple collection names in one prompt
TrendingTop collections ranked by a chosen metric. Default count is 5, max is 10."top 5 by volume", "trending", "highest floor price"
PortfolioNFTs held in a specific wallet address. Triggered automatically by any 0x address.Any 0x wallet address in the prompt
Top SalesHighest sales from a single collection over a timeframe (24h / 7d / 30d / all-time, default 30d)."top sales", "highest sales last 7 days", "most expensive sales"
ShowcaseDisplay collection logos only, no data or metrics. Pure visual gallery."show these logos", "exhibit on a wall", "put on board" without metric keywords
Art styles
AutoDefault. With a PFP attached the whole scene matches your character’s style; without one it renders as a hand-drawn editorial illustration.
Anime / MangaJapanese illustration style with clean lines and vibrant colors.
3DStylised 3D render look. Smooth surfaces, soft lighting.
PixelRetro pixel-art aesthetic. Best for gaming and 8-bit themes.
WatercolorSoft painted look with paper texture and bleed effects.
CartoonBold outlines and flat colors. Approachable, playful.
MinimalistClean shapes, limited palette, maximum negative space.
SketchHand-drawn pencil or charcoal look with visible linework.
Quality and aspect ratios
Quality
Credits
When to use
1K
3
Quick iteration, social posts
2K
4
High-detail finals, prints up to A5
4K
6
Maximum detail, archival quality
Aspect ratios available: 1:1 (Square), 16:9 (Landscape), 9:16 (Portrait), 4:3, 3:4. Use 1:1 or 16:9 for X/Twitter, 9:16 for Stories and TikTok.
Edit any scene
Finished scenes can be revised without starting over: use the pencil on the card (Edit in the card menu on mobile) or the Edit button in the fullscreen view, describe the change, and a new card is generated alongside the original. Ratio defaults to Auto, the edit keeps the source aspect unless you pick one, and the cost matches a single scene at the selected quality.
Collection names and wallet addresses are recognised automatically. Mention them anywhere in your prompt and the pipeline picks them up.
Tier List
The Tier List tool at /tier-list is a community ranking builder. Create S-to-D tier lists using NFT collection images or your own uploads, publish them to the gallery, and react to what others have ranked.
Building a tier list
1Go to /tier-list and click "Create Tier List" to open the builder.
2Add items to the unranked pool via the sidebar: upload image files, paste from clipboard, or search any OpenSea collection by name.
3Drag items from the unranked pool into a tier row (S, A, B, C, D by default). Items can be reordered within a tier or moved between tiers at any time.
4Customise tier labels, colors, slot shape (sharp / rounded / circle), and whether item names appear below each slot.
5Set a title and adjust typography for the title and tier labels using the toolbar font pickers.
6Optionally upload a background image. Use the settings popover to control opacity, blur, and dark overlay.
Managing tiers
Add a tierClick the "+ Add" button in the Tiers sidebar section. New tiers get the next color in the cycle automatically.
Rename a tierClick the label input in the sidebar and type. Labels support up to 3 lines and 8 characters per line.
Reorder tiersUse the up / down arrow buttons next to each tier in the sidebar.
Delete a tierClick the trash icon next to the tier. Items inside are removed along with it.
Change tier colorClick the color dot to the left of the tier label input.
Adding items
File uploadClick the Upload section in the sidebar or the + button inside any tier slot. Multiple files can be selected at once for the unranked pool.Sidebar "Upload" section or in-slot + button
Clipboard pasteCopy any image anywhere and paste (Ctrl/⌘ + V) on the builder page. The image lands in the unranked pool.Ctrl / ⌘ + V
Drag files inDrop image files directly onto the sidebar area. All dropped files go to the unranked pool.Native file drag-and-drop onto sidebar
OpenSea searchType a collection name in the OpenSea Search section of the sidebar. Click a result to add its collection image to the unranked pool.Sidebar "OpenSea Search" section
Exporting
Click "Export PNG" in the toolbar to download a 1920×1080 PNG of your tier list. The export uses a pure Canvas 2D renderer, no watermarks, no browser rendering quirks.
The builder auto-saves to your browser's local storage on every change. Your tier list is restored automatically if you refresh or close the tab.
Publishing to the gallery
1Click "Publish" in the toolbar. A modal opens.
2Confirm or edit the title (up to 80 characters). Add an optional description to give context about what you ranked.
3Click Publish. The system uploads any local images to cloud storage, generates a thumbnail, and adds your tier list to the community gallery.
4After publishing, click "View in Gallery" to see your list live, or "Keep Editing" to continue working.
Publishing requires a ReelyArt account. It uses no credits, the Tier List builder is completely free.
Community gallery
The gallery at /tier-list shows all published tier lists. Two tabs are available.
Community tabAll published tier lists from every user. Search by title or description, sort by Most Recent, Most 🔥, Most ❤️, or Most Viewed, and filter by time period.
My Lists tabYour own published tier lists. Click any card to reopen it in the builder for editing or exporting.
Viewing and editing a published list
Click any card in the gallery to open that tier list in the builder. You can edit, re-rank, and export it as a PNG. If the list is yours, you can publish an updated version.
Reactions
🔥 FireReact with fire to mark a tier list as based or agree with the ranking. Click again to remove your reaction.
❤️ HeartReact with heart to save a tier list as a favourite. Click again to remove.
Sign in requiredReactions require a ReelyArt account. The count updates instantly (optimistic) and syncs with the server.
Communities
Every verified community gets its own page on ReelyArt. Open one from /community and you get the project’s press kit, a chat bar to generate with, a feed of everything members have published, and your own private creations tab. New accounts get 3 free generations to spend across any community, no card, no subscription.
1Open /community and pick a project, or go straight to its page at /community/<name>.
2Type what you want in the chat bar at the bottom. Pick a model (fourteen to choose from, twelve of which a free plan covers at their opening tier), an aspect ratio, quality, and how many variations to generate.
3Optional: click a press kit asset to push the project’s official artwork into the chat as a reference, or attach your own image with the paperclip.
4Results land in My creations. Press Publish on one to share it with the community feed.
Free generations
3 per accountYour free generations are counted once per account, not per community, spend them wherever you like. They cover the twelve models a free plan can run, at their opening tier.
Then creditsOnce they run out, generating on a community page costs the same credits as the equivalent Studio generation. Failed generations are always refunded.
Extra grantsWe occasionally hand a community extra free generations for a campaign. If you have any, the counter above the chat bar shows them.
Publishing
Anything you madePublish from a community page, from My creations, or from your Library: pick the community, and the image is copied into its feed.
Prompt is optionalChoose whether your prompt is shown with the post. Keep it private and only the image is shared.
Use this templateIf a post was made from a Studio template, its name appears on the card and anyone can open the same template and make their own version.
ModerationA project can hide a post from its own feed at any time; only ReelyArt deletes permanently. You can always delete your own post.
Getting a page for your project
Pages are opened manually, one at a time. Apply at /community/apply with your project and links. We talk to the team first, then hand the page over. Owners edit their own name, description, logo, banner, links, and press kit (up to 20 JPG or PNG assets, 10MB each) and moderate their feed.
Press kit assets are the fastest way to keep a community’s output on-brand: one click puts the official character or logo into the chat as a reference before you write a word.
Prompt Guide
A good prompt is short, specific, and visual. You do not need to be a designer or a copywriter. A little structure goes a long way.
The five-part formula
Subject + Action or Setting + Style + Mood + Details. Most prompts need 1 to 3 sentences, not paragraphs.
Part
What it controls
Subject
What the image is about: the character, product, event, or collection.
Action / Setting
The specific event, context, or scene.
Style
Visual aesthetic: minimal, retro, cyberpunk, watercolor, etc.
These are custom prompts you can add on top of a Studio template to nudge the result. For most standard templates an empty prompt works best.
Standard template tweakdarker tones, slightly more dramatic lighting, keep the subject's expression neutral.
Custom modeA neon-lit ramen shop in Tokyo at midnight, steam rising from the bowls, cinematic 35mm look, shallow depth of field.
Header Studio prompt examples
X HeaderCyberpunk Tokyo skyline at night, neon reflections on wet streets, deep purple and electric blue palette. Leave the left third clean for an avatar.
Discord BannerMinimal abstract waves in slate gray and white, architectural composition, generous negative space.
Content prompt examples
GIF (with reference image)slowly breathing, eyes glowing blue, subtle hair movement, gentle camera push in.
TrendingShow top 5 trending collections by 7-day volume on a blackboard. Sketch style.
ComparisonCompare Azuki and Pudgy Penguins floor prices on a newspaper front page. Watercolor style.
ShowcaseDisplay BAYC, CryptoPunks, Doodles, and Milady on a gallery wall. Minimalist style.
PortfolioShow top 5 NFTs in 0xa8e6b51519e7f0ef973a4a0c58b8ce61343ac147 on a cork board. Title "Diamond Hands". 3D style.
Color overrides
Any color word in your prompt overrides the selected style. The AI will always prioritise your explicit colors.
"orange tones with dark background"Forces an orange palette regardless of the selected style.
"soft pastels"Forces a pastel palette even if Minimalist is selected.
"monochrome black and white"Forces a grayscale palette.
What not to do
Do not write 500-word promptsThe AI performs better with 1 to 3 focused sentences. Long prompts dilute attention.
Do not contradict yourselfEither describe the look in the prompt or pick a style. Not both with conflicting directions.
Do not reference images you have not attached"@image1" only works after you upload at least one image.
Do not use absolute coordinatesAvoid "top-left corner, exactly 50px from edge". Describe the composition in words instead, for example "leave the left third clear".
Tips & Tricks
Small habits that compound. These are the shortcuts power users learn after a few days.
Paste from clipboardCopy any image from a browser, screenshot tool, or another app and paste it directly into any Studio, Content, or Lab upload zone.
Revise, do not restartAfter generating, ask for changes instead of starting over: "make the title larger and change it to dark red". Studio, Header Studio, Lab, and Content Studio meme results have the same power via the pencil Edit button on any card.
Library is your archiveEvery Studio image, header, Content asset, and Lab scene lives in /library. Search and rerun any past creation.
Style preset vs prompt colorsPrompt colors always win. Use the preset for the base mood, then override individual colors in the prompt.
Bulk download in Content StudioSelect multiple items in your gallery and download them as a batch.
FAQ
The questions we hear most often. If yours is not here, the feedback button in the footer goes straight to our inbox.
Credits were deducted but the generation failedFailed generations refund credits automatically within 30 seconds. If your balance still looks off after a refresh, send us feedback with the prompt text.
My OpenSea collection was not foundLab matches collection slugs from OpenSea. If the AI guesses the wrong slug, search by the exact slug instead (for example "boredapeyachtclub" rather than "BAYC").
I need a vertical or horizontal GIFChoose the ratio in the GIF composer before generating: square 1:1, portrait 3:4 or landscape 4:3. The scene, the video and the GIF are made in that shape, so nothing is cropped, and the price is the same for every ratio. Gallery cards stay square and show the whole GIF inside; the preview shows its real shape.
My GIF is too shortThe length slider under the model picker runs from 4 to 8 seconds. The price scales with it (seconds divided by 4, rounded to whole credits) and anything above 4 seconds needs a paid plan. Longer clips also make heavier GIFs, so the encoder trades some detail to stay inside the 8 MB limit.
Can I edit an image I already generated?Yes. In Studio, Header Studio, and Lab every finished image has an Edit action, the pencil on the card or the Edit button in the fullscreen view. Describe the change in plain language and a revised copy is generated as a new gallery card; the original stays untouched. An edit costs the same as a single generation at the quality you select. Meme results in Content Studio can be edited the same way for a flat 2 credits.
Subscription vs pack creditsSubscription credits reset monthly and do not carry over. Pack credits never expire. The system always spends pack credits first, then subscription credits.
Can I upgrade or downgrade my plan?Yes. Upgrades take effect immediately. You pay only the difference between the new plan price and the value of your unused subscription credits, charged as a one-time invoice. Monthly downgrades take effect at the end of your current billing period. Annual plan changes (including switching to monthly) are not available mid-year; they apply at the next renewal date. Monthly subscribers can switch to any annual plan at any time.
Can I use ReelyArt outputs commerciallyYes. You own the outputs you generate. Review /terms for the full licensing details.
How do I cancel my subscription?Open your Account page and click Cancel Plan. Cancellation takes effect at the end of your current billing period, your access, features, and credits remain fully active until then. Annual subscriptions stop renewing at the end of the paid 12-month period. No partial refund is issued for remaining unused time.
My renewal payment failed. What now?Stripe retries the charge automatically over the following days and emails you a link to update your payment method. Unused subscription credits from the ended cycle expire when the renewal payment fails; as soon as a retry succeeds, your full monthly allocation is granted again. If every retry fails, the subscription is cancelled. Permanent credits are never affected.
How do I delete my account?Go to your Account page and scroll to the bottom. Click Delete, type "delete" to confirm, then click Delete My Account. This permanently removes your profile, all designs, generated content, credits, and subscription data. The action is immediate and cannot be undone.
Ready to start creating?
Sign up with Google or X and 10 free credits are waiting. No credit card required.
