Instant Studio · Just make something
Every top model, one prompt away — make it now
Pick any state-of-the-art image or video model, describe a scene, and generate — no brand kit, no setup. Results stream into a live feed while you start the next, everything auto-saves to My Generations, and your own creations feed the next shot. The fast "just make something right now" front door — or ask Claude to run it for you.
Every top model · one place
The full SOTA roster — image and video
Pick by the job — the tag on each model tells you what it’s best at. Click any model to open it in Create & Generate, pre-selected and ready for a prompt. New SOTA models land here as they ship.
The flywheel · Every generation feeds the next
Pick a model, describe it, make it — then make more
Choose any top model, type a scene, and generate — no brand kit needed. A set streams into a live feed while you start the next, everything auto-saves to My Generations, and your own creations feed the next shot — as references, or as a video's element / start frame / end frame.
Pick a model · describe it
“a matte-black energy drink can, splash of citrus, dramatic light”
The live feed
images stream in · start another anytime
My Generations
Generate in parallel
Sets fire without blocking — start the next idea while the first is still coming in. The feed fills as each finishes.
Auto-saved history
Everything lands in My Generations automatically. Your session is never lost, and it's all there to reuse.
Your work is your input
Pull any past generation back in — as a reference for an image model, or an element / start frame / end frame for a video model.
How it works, start to finish
Pick a model, describe it, generate — a live feed that auto-saves, saves the keepers on hover, and feeds your work back in as references and frames.
The Models
Every top model in one place — image and video. Pick by the job: photoreal, transparent cut-outs, typography, cinematic motion. A showcase previews each model's look before you commit, and swaps as you switch.
Describe it
Type the scene you imagine — a product, a character, a mood, a style. No brand kit, no setup. Just describe and generate.
Generate — and keep going
Fire a set and it streams in without blocking the page. Kick off another the moment the first is underway — images land fast, video renders in the background.
The Feed
Everything you make stacks into a live feed, newest first — a running roll of the session's ideas you can scroll, compare, and act on.
Auto-saved
Nothing is ever lost — every generation auto-saves to your Albums under My Generations, so you can step away and it's all still there.
Save the keepers
Hover any result for the save toolbar: keep it as a brand element, drop it on a board or a brief, save to an album, edit it, or enlarge. The ones you love become reusable across StyleForge.
Polish in the editor
Open any image or video in the built-in editor — recrop, rotate, flip, scale, and fine-tune sharpness and tone, right in the app. It saves a fresh copy into the feed and My Generations, leaving your original untouched.
Reference images
Add reference images to image-to-image models to steer a generation — bring a look or a subject to match, right from the prompt bar.
Your work becomes your input
Open My Generations and pull any past creation back in — as a reference for an image model, or as an element / start frame / end frame for a video model. Every generation can feed the next.
Or just ask
Describe it — Claude generates it
The exact same generation runs from a conversation. Claude fires a set with the model you name and hands back a live preview link with numbered results — then “change #2” iterates in place, and it can animate a still or save a keeper to your kit. It states the price and asks before spending.
“Make me 4 shots of a matte-black energy drink can with a citrus splash, on Nano Banana Pro.”
Claude fires the set → a live preview link with numbered results → “change #2” to iterate, or “turn #2 into a Veo 3 video.”
The live feed
numbered results · iterate by talking
Ask Claude to
Make it from a conversation
“Make me 4 hero shots of a matte-black energy drink can with a splash of citrus, on Nano Banana Pro.”
Iterate a variation in place
“Change #2 — deep navy background, more splash.”
Animate & save
“Turn #2 into an 8-second Veo 3 video — gentle motion and ambient sound.”
New to this? Connect your agent and see everything you can ask →
Everything in Create
The full SOTA roster, parallel generation, a save toolbar on every result, and your own library steering the next shot — all kit-agnostic, all live.
Instant
Every SOTA model
Image: Nano Banana Pro, GPT Image 2 / 1.5, Seedream 4.5 / 4 / 5 Lite, FLUX 2 Pro / 1.1 Pro / Ultra, Qwen, Ideogram v3, Lucid Origin, Recraft v4 Pro / v4 / v3. Video: Kling Omni / v3 / v2.6, Veo 3 / Fast, Seedance 2 / 2 Omni / 1 Pro, Hailuo 02, WAN 2.2, Grok Imagine. The controls fit whichever you pick.
No setup
Kit-agnostic
No brand kit required — the fast 'just make something' door. When you want on-brand, campaign-ready production, the same models live inside a brand kit.
Flow
Generate while you generate
Sets fire non-blocking — start another the moment the first is underway. The feed fills in as each finishes; you never sit on a spinner.
Keep
Auto-save + My Generations
Every result auto-saves to your Albums. The My Generations panel is your whole history — reopen, download, or feed any of it straight back into a new generation.
Save toolbar
One hover to keep it
Hover a result for the toolbar — save as a brand element, add to a board or brief, save to an album, edit, or enlarge. Keepers become reusable brand assets.
Refine
Built-in image & video editor
Recrop, rotate, flip, scale, and adjust sharpness and tone on any image or video, right in the app — no round-trip to another tool. Every edit saves as a new copy to the feed and My Generations; the original is never overwritten.
Steer
References, elements & frames
Drop reference images into image-to-image models; or pull a generation from My Generations as a video element, a start frame, or an end frame. Your own library steers the next shot.
Any format
Aspect, count & duration
Pick the aspect and how many per run; video adds a duration. Each model exposes exactly the controls it really supports — nothing that won't work.
Create is the quick 'make something now' front door — non-persistent, no kit needed. Results auto-save to your Albums; you save the keepers you love deliberately.
Images are cheap and near-instant; video is slower (a few minutes, sometimes more) and costs more. It renders in the background and drops into the feed when it's ready.
Most of the newer video models come back with audio — Kling Omni, v3 and v2.6, Veo 3 and Veo 3 Fast, Seedance 2 and Seedance 2 Omni, and Grok Imagine. Seedance 1 Pro, Hailuo 02 and WAN 2.2 are silent by design. The model showcase only offers a Sound button on clips that actually have audio to hear.