# FramePack Studio Image Generator

Choose a current image model, use only the prompt, reference, and output controls shown by its form, then inspect the returned image.

> Canonical: https://framepack.studio/ai-image-generator

## How the FramePack Studio Image Generator Works

Two entry points - a written description, or a picture you already have - and a loop for choosing between options rather than polishing one.

### Start from a written prompt

Describe the subject, the setting and the light in that order. Each one narrows what the next word can mean.

### Or start from an image you have

Image-to-image holds composition and colour steady while the style, material or subject treatment changes.

### Generate several options at once

A grid of variations surfaces directions you would never have thought to describe. It beats refining a single image.

### Change the prompt, not the pixels

Adjusting one word about lens, light or material usually moves the result further than editing the output by hand.

## Three Ways People Start an Image

Almost every session begins in one of three places, and each one wants a different first move.

### From a description alone

You know what you want and have no reference. Write the subject first, then the setting, then the lighting - working outward, so each choice constrains the next instead of fighting it.

### From a reference image

You have a composition that works and want a different palette, era or subject inside it. Attach the reference and describe only what should change; everything you leave unsaid stays put.

### From something you just generated

The direction is right and one element is wrong. Feed the result back in and move a single attribute, so you can tell which change actually did the work.

## Why Generate Images in the Browser

No local environment to build, and nothing to keep up to date between sessions.

### Nothing to install

No Python environment, no driver versions, and no multi-gigabyte checkpoint download before the first image.

### The same result on any machine

Generation runs server-side, so a laptop returns what a workstation returns for the same prompt.

### Iterate wherever you are

Open a tab and keep the good ones. The session lives with your account, not in a folder on one computer.

### Carry a still into motion

An image made here can become the first frame of a video without exporting and re-uploading it.

## How to Generate an Image

Four steps from an idea to a set of options worth choosing between.

### Describe the subject and the frame

Say what is in the picture and how it is composed before reaching for style words.

### Attach a reference when layout matters

Image-to-image keeps composition and colour fixed while the rest is free to change.

### Generate a batch and compare

View candidates together. Side by side, the strongest direction is usually obvious within seconds.

### Re-run the winner with one change

Move a single attribute so you can attribute the improvement to something specific.

## What the Image Workspace Includes

The parts of the studio that matter when you are making stills.

### Text-to-image and image-to-image

Both entry points in one place, so switching between them costs nothing.

### Server-side rendering

The work happens remotely; your machine only displays the result.

### Side-by-side comparison

Earlier attempts stay visible instead of being replaced by the newest one.

### Files at working resolution

Download straight into a layout, a deck, or a video timeline.

## What a run costs and returns

| Measure | Detail |
| --- | --- |
| 17 models | Available in the form today, each priced in the same credits. |
| 50–150 credits | Per run, depending on the model and length chosen. |
| Every paid plan | Includes a commercial licence for what you make. |
| 1,800 characters of prompt | The longest prompt every model on this page accepts. |

## Image Generation Questions

What people ask before their first image.

### What is the difference between text-to-image and image-to-image?

Text-to-image builds a picture from a description alone. Image-to-image starts from a picture you supply and changes it according to your prompt, which keeps composition far more stable than describing that composition in words.

### Do I need to install anything?

No. Generation runs in the browser against a hosted model, so there is no local environment to build, no checkpoint to download, and nothing to update between sessions.

### Why do two runs of the same prompt look different?

Image models sample randomly unless the seed is held fixed. Re-running a prompt explores the space around it rather than reproducing one result, which is useful early and frustrating late.

### How do I keep a character or product consistent?

Supply a reference image and change one attribute at a time. Describing consistency in words alone is far less reliable than showing the model what to hold on to.

### Can I use a generated image in a video?

Yes. A still generated here can be used as the first frame for image-to-video, which gives you more control over the opening shot than describing it in text.

### How many credits does FramePack Studio Image Generator use?

FramePack Studio Image Generator draws between 50 and 150 credits per run, depending on the model and length chosen in the form.

### What limits should I check before using FramePack Studio Image Generator?

Generated output from FramePack Studio Image Generator can contain visual, audio, or factual errors. Review each result before publishing, and upload only material you are authorized to use.

### What happens to the files I upload to FramePack Studio Image Generator?

An upload is stored against your own account, not pooled with other people's. Deleting the account removes it: the control is in Settings under Security, the request is confirmed by email, and the erasure removes the stored objects it owns. How long anything else is kept, and which processors handle it, is stated in the privacy policy.

### Can I use what FramePack Studio Image Generator makes commercially?

Every paid plan includes a commercial licence for what FramePack Studio Image Generator makes; the free tier is for personal use.

### How does FramePack Studio Image Generator compare, and what alternatives fit a different workflow?

FramePack Studio Image Generator: FramePack Studio image generator: choose a current image model, use its prompt or reference inputs, review the credit estimate, and inspect the result. Related options are FramePack Studio Music Generator: FramePack Studio music generator exposes prompt and custom music workflows, a visible credit estimate, and a returned audio result to review before use, FramePack Studio Video Generator: FramePack Studio video generator: choose a current video model, use its text or image inputs, review the credit estimate, and inspect the returned clip, FramePack AI - an independent hosted workflow guide: FramePack AI on this domain is an independently operated workflow guide that separates the browser generator from the upstream research repository, and FramePack Studio: FramePack Studio is an independent workflow guide around FramePack: understand the project boundary, plan a run and continue to the current browser generator. Choose by the required input, delivered output, controls, limits, current availability, and credit cost; the related-tools section links each canonical page.

## Generate Your First Image

Start from a prompt or a reference and see what comes back.

## Related tools

### [FramePack Studio Music Generator](https://framepack.studio/ai-music-generator)

### [FramePack Studio Video Generator](https://framepack.studio/ai-video-generator)

_Last updated July 21, 2026_