Muse Image Generator
Muse Image generator online: Meta's first image model, square text-to-image at 30 credits, free to try.
Sample images
A square retro travel poster titled 'VIDLARK' in bold lime-yellow (#EBF958) letters, a small fuzzy lime-green felt bird with a round black eye flying over layered sunset mountains, flat screen-print style, warm paper texture, the words 'Make it move' along the bottom.
Muse Image is the first image model from Meta Superintelligence Labs, announced on July 7, 2026 and built into Meta AI, where Meta says it plans the layout before drawing and renders legible text. On Vidlark it does one thing: text to image as a square 1:1 picture at 1K, 30 credits each. There is no photo editing and no other aspect ratio here. Free accounts can use it with the sign-up credits.
Updated
At a glance
- Made by
- Meta
- Price
- From 30 credits per image
- Size
- 1K
- Input
- Text
- Per run
- 1–4 images
- Available on
- Every account, free included
- Account
- Free account required, 100 credits on sign-up, no card
- Unlimited plan
- Free (fair use)
Vidlark is an independent platform and isn't affiliated with Meta.
How to use Muse Image
- 1
Describe the image
Type a prompt.
- 2
Pick the settings
1K; the price shows before you start.
- 3
Generate and download
Download the images when they are ready.
What it does well
Plans the picture first
Meta built Muse Image to work out the layout before it draws, so prompts that place several things, like a poster or a how-to card, come out arranged as described.
Text you can read
Short headlines, labels and captions inside the image come out spelled as written, which suits posters, menus and social graphics.
Square and quick
Every result is a 1:1 image at 1K for 30 credits, the size feeds and profile pictures use, so drafts are cheap to repeat.
Settings and prices
Each image has a fixed price; the button always shows the exact total before you generate.
| Setting | Price | Example | Unlimited plan |
|---|---|---|---|
| 1K | 30 credits / image | 4 images = 120 credits | Free |
Prompt tips for better images
Compose for a square
Put the subject in the centre and say what fills the corners; a wide landscape squeezed into 1:1 loses its edges.
Quote the words
Put any text in quotes and keep it short: 'OPEN LATE' or a three-word headline renders more reliably than a paragraph.
Run several, keep one
At 30 credits a picture, make a few versions of the same prompt and pick the best rather than rewriting it each time.
Example prompt
A ceramic coffee cup on a linen tablecloth by a window, soft morning light, shallow depth of field, 85mm product photo.
Use this promptWhat people make with it
01
Social posts
Square visuals for Instagram, Facebook and WhatsApp channels: a product on a clean background, a quote card, a small illustrated scene.
02
Avatars and stickers
Characters, mascots and icons that sit centred in a square, ready to crop into a circle or cut out as a sticker.
03
Comparing models on one prompt
Run the same prompt here and on Nano Banana 2.1 or GPT Image 2 to see which look fits before you make the final.
How it compares
How it lines up against similar image models on this site; every one uses the same credit balance.
| Model | Best for | Top quality | Per run | Edit photos | From |
|---|---|---|---|---|---|
| Muse Image | Cheap square images with readable text, from a prompt. | 1K | 1–4 | No | 30 credits |
| Nano Banana 2.1 | Images with readable text and detailed edits, up to 4K. | 4K | 1–4 | Yes | 40 credits |
| GPT Image 2 | The previous GPT Image look, with 8 references. | 4K | 1–4 | Yes | 30 credits |
Other models to try
All models
GPT Image 2.5 FlareUnlimitedOpenAI's GPT Image 2.5, balanced and fast: accurate text and precise edits.from 40 creditsGPT Image 2.5 SunburstUnlimitedThe high-precision GPT Image 2.5 for detailed commercial images.from 40 credits
Nano Banana ProUp to 8 reference images for consistent characters.from 55 credits· Paid plansNano Banana 2UnlimitedFlash speed with Pro-level consistency.from 30 credits
Seedream 5 ProByteDance's photo-real images up to 2K.from 45 credits· Paid plansSeedream 5 FlashUnlimitedFast, cheap Seedream for drafts and variations.from 30 creditsMuse Image questions
Is there a free website to try Meta Muse Image?
Meta's own Meta AI app and meta.ai include Muse Image free for everyday use, with paid plans for more. On Vidlark you can run it in the browser with the 100 sign-up credits and no card: one square image is 30 credits.
Can Muse Image edit my photos here?
Not on Vidlark. Here Muse Image makes new pictures from a text prompt only. Meta's photo editing, blending and object removal run inside the Meta AI app; for edits on Vidlark, pick Nano Banana 2.1 or GPT Image 2.
What sizes does Muse Image make on Vidlark?
One size: a square 1:1 image at 1K, 30 credits each. For wide or tall pictures use another model.
Who makes Muse Image?
Meta Superintelligence Labs, the AI group at Meta; it is their first image model, announced on July 7, 2026. Vidlark runs the model through an API provider and is not affiliated with or endorsed by Meta.
What sizes can it make?
1K, in square, portrait or landscape. Each run makes 1–4 images.
Credits, refunds, watermarks and commercial use work the same on every model. See the pricing FAQ
Make a Muse Image picture
100 free credits when you sign up. Failed runs are refunded.