Type a prompt, pick a model like FLUX 3, Seedream or Nano Banana, and generate. Turn a premise into a cast and a storyboard with Story. When you want more control, switch to Node mode and wire prompts, models, masks and video into pipelines you can rerun.
Free plan with daily credits. Pro from $12 a month.
Where everyone starts. One prompt, one click.
Prompt
Underwater portrait, backlit hair drifting, bubbles rising, deep blue
ModelFLUX.2 Pro ⌄
Aspect4:5 ⌄
Style
Film · CinematicToon · WatercolorAbstract · Surreal
600 creditsGenerate
Made in Fluxion
Nothing on this wall was photographed.
Every frame here came out of a node pipeline on the Fluxion canvas.
How it works
Start simple. Go deeper when you want.
Simple mode
For everyone, and where you start.
One panel, no wiring. It's the default on desktop and the whole app on your phone. A Story tab sits beside it when you want to tell one.
1Describe itType what you want to see, or start from a style preset like Cinematic or Watercolor.
2Pick a modelFLUX 3, Seedream, GPT Image, Nano Banana and more. The credit cost shows before you generate.
3GenerateSketch four quick drafts first if you like. The result saves to your library.
Node mode · advanced
When you want control, wire it up.
Flip the switch to Node and every step becomes a node on a canvas. Connect a prompt to a model, a model to a mask, a mask to an inpaint. Tweak one node and rerun only what changed.
1
Write
Text PromptWrite it yourself
AI PromptTurn a rough idea into a detailed prompt
StoryA premise becomes a cast and a shot list
2
Generate
Image Diffusion13 image models in one node
DraftsFour quick sketches, then finalize one
ReferencesUp to 14 images to guide or edit
3
Refine
MaskPaint the area to change
InpaintingFill the mask from a prompt
4
Animate and keep
Video DiffusionText or image to video, with sound
EditCut clips together, with transitions and AI music
Story · new
From one line to a storyboard.
Give Story a premise and it writes the rest: who's in it, where it happens, how it looks, and every shot. Then render it frame by frame with the same faces and places throughout. It's a tab in Simple mode and a node in Node mode.
1
PremiseOne line is enough. Pick how many shots, and a style if you have one in mind.
2
Story scriptA title, a look, a cast, its places and a numbered shot list, all editable. Rewrite any shot, or the whole cast.
3
Cast and placesA portrait for each character, or cast a real person from a photo; a set plate for each place. Every frame draws them the same.
4
StoryboardRender it shot by shot with any model, then refine the keepers with a stronger one. Edit any frame’s prompt, redo, reorder, export a PDF.
5
AnimateTurn each frame into a clip with Seedance, Kling, Veo, LTX and more, then play the story through and export it as one video.
The Far-Side Message5 shots
Premise: A lighthouse keeper finds a message in a bottle, addressed to her and dated fifty years from now.
35mm film still, soft grainslate blue, weathered gray, warm amber
Cast · Seedream 5 Lite · 700 each
1Marathe lighthouse keeper
2Pipher terrier
Places · set plates
The lighthouse shoreSeedream 5 Lite · 700The lantern roomNano Banana 2 · 1,340
1Extreme wide
The bottle washes upNano Banana 2 · 1,340
2InsertClip
Her own face, fifty years onGPT Image 2 · 940 · then LTX-2.5 · 5 s · 1,500
3Over the shoulder
Then and now, in the glassNano Banana 2 · 1,340
4Top-down
The portrait goes back inGPT Image 2 · 940
5Wide
The message sent out to seaGPT Image 2 · 940
Every frame as it was rendered, with its model and what it cost in credits. Export a storyboard or a comic page, or animate every frame and play it through.
Edit · new
Clips in, a finished video out.
Cut your clips together without leaving Fluxion: trim them, put them in order, join them with crossfades, and score the whole thing with AI music. It's a tab in Simple mode and a node in Node mode.
1
Pick your clipsSelect videos in your gallery and choose Add to edit, or drag them in. Story clips, Video Diffusion runs and your own uploads all work.
2
Trim and orderDrag the handles to cut each clip down, reorder them, and use a clip twice. The file is never cut, so you can always change your mind.
3
Transitions and fadesA cut, a crossfade or a dip to black between any two clips, and a fade in and out for the whole video. Mute any clip’s own sound.
4
AI musicDescribe the mood and make a track with ElevenLabs Music, sized to your edit, or upload your own. Keep every take, and use up to three tracks that crossfade from one to the next. About 3,740 credits for 30 seconds.
5
ExportOne mp4 at 720p or 1080p in any shape, or just the soundtrack as audio, free. It lands in your gallery, and My edits keeps every cut to come back to.
The Orchard Keeps TimeTwo musicians, an orchard at dusk, one song. Written with Story, cut and scored in Edit.
Cast · Seedream 5 Lite · 700 each
The dulcimer playerThe sampler player
Frames, then clips · in cut order
GPT Image 2 · 940then LTX-2.5 · 5 s · 1,500GPT Image 2 · 940then FLUX 3 Video · 5 s · 12,750GPT Image 2 · 940then LTX-2.5 · 5 s · 1,500GPT Image 2 · 940then LTX-2.5 · 5 s · 1,500
Music: Two ElevenLabs Music tracks, 18 s each, crossfading · 2,250 eachCut: 4 clips, a crossfade and a dip to black · 16 s at 720p · export free
Your edit5 clips · export free
Clips
1
2
3
4
5
Crossfade →Fade →Cut →Cut →Dip to black →Dip →Crossfade →Fade →
Music · AI or your own
Fade in 1sMusic 80%Fade out 2s
Export video · 1080p
Models
25 models. One canvas.
Each generation costs credits by model, shown on the node before you run it. New models show up here as they land.
Text to image
FLUX.2 KleinBlack Forest Labs
Fast drafts. Explore ideas before you commit.
From 10 credits
FLUX.2 ProBlack Forest Labs
The everyday workhorse: sharp detail, strong prompt following, up to 8 references.
From 450 credits
FLUX.2 FlexBlack Forest Labs
Max-quality generation and editing with up to 10 references. Great for typography.
From 600 credits
FLUX.2 MaxBlack Forest Labs
The highest-fidelity FLUX for final frames.
From 1,100 credits
FLUX 3NewBlack Forest Labs
BFL's newest model: generation and editing with up to 10 references, and optional web search to get real places and products right.
From 480 credits
Nano Banana 2Google
Conversational edits, multi-image fusion and character consistency, up to 14 references.
From 1,340 credits
Imagen 4 UltraGoogle
When quality matters most: fine detail, photorealism and clean text.
1,200 credits
GPT Image 2OpenAI
Follows instructions to the letter. Dense text, layouts, precise edits, transparent PNGs.
From 240 credits
Seedream 5 LiteByteDance
Reasons about your prompt. Story sequences of up to 8 frames, and example-based editing with up to 14 references.
From 700 credits
Photon FlashLuma
Fast, cinematic compositions. One reference for a character, a style or a layout.
200 credits
PhotonLuma
Luma’s full model for richer light and texture. One reference for a character, a style or a layout.
600 credits
FLUX Kontext ProBlack Forest Labs
Combines two images with one instruction ("put the jacket from image 2 on the person in image 1").
800 credits
Recraft V4.1 ProRecraft
Design taste at ~2048 px: posters, brand visuals, illustration.
5,000 credits
Edit and combine
FLUX Fill DevBlack Forest Labs
Paint a mask, describe the change, inpaint.
800 credits
FLUX Fill ProBlack Forest Labs
Pro-quality inpainting and outpainting.
1,000 credits
Text or image to video
P-VideoPruna AI
Fast and inexpensive, with a draft mode for trying ideas. Text or image to video.
From 230 credits
Seedance 2.0 FastByteDance
Keeps characters consistent from up to 9 reference images, with sound. Quicker and cheaper than Seedance 2.0.
From 4,200 credits
Seedance 2.0ByteDance
ByteDance’s best video model: cinematic motion, consistent characters from references, sound, up to 1080p.
From 4,800 credits
Kling 3.0Kuaishou
Cinematic shots up to 15 seconds, from a start and end frame, with optional sound.
From 7,560 credits
Kling 3.0 OmniKuaishou
Kling with up to 7 reference images, for the same characters, objects or set in every shot. Optional sound.
From 7,560 credits
Veo 3.1 FastGoogle
The most lifelike motion and sound, in clips up to 8 seconds.
From 6,000 credits
FLUX 3 VideoNewBlack Forest Labs
BFL's first video model: clips up to 20 seconds with sound, from text, a start frame or a start and end frame. Draft mode for cheap previews.
From 4,500 credits
Wan 2.7Alibaba
Open model: animates an image, or moves between a first and last frame, with sound.
From 3,000 credits
LTX-2.5Lightricks
Open model on our own GPUs: quick and inexpensive, with sound. Text or image to video, 16:9 or 9:16.
From 900 credits
Writing · Story
GPT-6 LunaOpenAI
Writes the story script: title, look, cast and a shot list, from one line of premise.
~20 credits a story
Your first image is one prompt away.
Free plan with daily credits. Pro from $12 a month.
1Marathe lighthouse keeper
2Pipher terrier
1Extreme wide
3Over the shoulder
4Top-down
5Wide
