StudioLM is a free AI image generator, with a $0.01 image API

StudioLM — Sun, 04 Oct 2026 05:37:50 +0000

StudioLM is a free AI image generator and editor that runs in the browser. The studio is called Imagen. It does not ask for a credit card, and generations in the browser are not metered. The same first-party models are available from code at $0.01 per image.

The product site is studiolm.dev. Imagen is the browser app. The API base URL is https://api.studiolm.dev. Docs are at docs.studiolm.dev.

What you do in the browser

Imagen is built for still images, not as a catalog of every model on the internet.

You describe a picture with tags, not a paragraph. A prompt is a comma-separated list of what belongs in the frame: subject, place, look. The site's own example is fox, mage, glowing runes, mossy forest.

You pick a look (anime, photoreal, anthro, or illustration) and a frame (square, portrait, or landscape). If part of the picture is wrong, you paint a mask and tag what should be there. Only that area is redrawn. The editor has layers, a brush, and zoom.

Finished images go into a library, sorted into boards. The community feed lets someone reuse another person's prompt and model in one tap.

A few other things the site states plainly:

A reference photo can steer a generation. Precise Reference is free for the first 10 photos each day.
A still can become a short clip, on desktop.
Style presets are one tap.
There is a phone layout and a desktop layout. It is not the desktop site squeezed onto a phone.
A Learn tab has short video guides inside the app.
Sign-up is email or Google. No card is required.

The API

Developers use a prepaid REST API. An image costs $0.01. Chat has a free daily token allowance. Credits are added in the dashboard with a card (Stripe) or crypto.

Routes follow the OpenAI layout:

POST /v1/images/generations generates an image
GET /v1/models lists models
POST /v1/chat/completions is chat
POST /v1/completions is text completions

There is a Python SDK: pip install studiolm. An API key is created from the dashboard after you log in at studiolm.dev. Image calls need prepaid credits. Imagen stays free.

A minimal image call, from the docs:

curl https://api.studiolm.dev/v1/images/generations \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{"prompt":"a red fox in a forest","aspect_ratio":"square"}'

The quickstart also shows chat on gemma-3-12b-it-qat.

Models

The public image API lists four models: AnythingV2, RealisticV1, AnthroV2, and AnthroV3.

AnthroNAI is a first-party model in Imagen only. It is not on the public API, and the API rejects it.

Text models are Gemma-class and show up on GET /v1/models.

The catalog is short on purpose. StudioLM runs those models itself. It does not rent you a GPU by the second, and it does not sell a Pro image plan.

StudioLM, Replicate, and Fal.ai

Replicate and Fal.ai are marketplaces and serverless GPU platforms. They host thousands of community and vendor models, including video and audio. StudioLM is a still-image product: a free browser studio plus a one-cent API on a few models it operates.

Use StudioLM when you want stills and you do not want to pick a checkpoint or pay per second. Use Replicate or Fal.ai when you need a specific catalog model, or a video or audio pipeline.

Pricing

Imagen in the browser is free. No credit card to start. Generations do not cost credits. The site says donations from people who use it cover the GPU bill.
The public image API is $0.01 USD per image, prepaid.
There is no StudioLM Pro image plan and no per-second GPU bill.

Questions the site already answers

Is it actually free? Yes, in the browser. Only the developer API costs money.

Can I edit one part of an image? Yes. Mask it, describe the replacement, and only that region is regenerated.

Does it work on a phone? Yes. Imagen runs in the browser with its own phone layout.

Do I need an account? Yes, a free one. Email or Google.

Where to start

Open Imagen from studiolm.dev if you want to make pictures. Read the API quickstart if you want the same models from code.

DEV Community: StudioLM