Which LoRAs work here?

LoRAs trained for FLUX.2 [klein] 9B, including those trained on its Base build: in our tests a LoRA trained on Base for the stronger levels loaded on this distilled model and changed the result as expected. FLUX.1 and klein 4B LoRAs ran without an error and changed nothing.

Should I use this or FLUX.2 Klein Base 9B LoRA?

This is the step-distilled build: renders in about ten seconds and holds a polished, photographic look. Base is the undistilled build its makers recommend for training; it took two to five times longer in our tests and gave softer, more varied pictures. The same LoRAs load on both, so you can train on Base and run here.

Can I use uncensored LoRAs for NSFW images?

Yes. LoRA files are not reviewed and the output is not filtered. Without a LoRA the model tends to cover or crop at the Strong level and tone down the Very strong level; LoRAs trained for that content on klein 9B removed that in our tests.

Why did my LoRA request fail?

Usually the link needs a login (gated Hugging Face files answer with HTTP 401) or points to the repository page instead of the file. Use a public URL that downloads the .safetensors file itself. Failed requests are not charged.

My LoRA loaded but nothing changed. Why?

Three causes in our tests: the file was trained for another model (FLUX.1 or klein 4B), it was saved in diffusers PEFT format (keys ending in .default.weight), or it is a style LoRA whose look the prompt already asks for. None of these raise an error. Compare the same seed with and without the file.

Each endpoint has one flat price per image. The aspect ratio, 1k or 2k, the number of edit sources and loading LoRAs do not change it. Failed requests are not charged.

FLUX.2 Klein 9B LoRA API: Uncensored NSFW Images, Your LoRAs

black-forest-labs/flux-2-klein-9b-lora/text-to-image

FLUX.2 Klein 9B LoRA text to image runs Black Forest Labs' step-distilled FLUX.2 [klein] 9B with up to three of your own LoRA files. Nine aspect ratios at 1k or 2k, about ten seconds per image, and no content filter.

From $0.018/image

Run model View API

Input

Text to image

Prompt523/5000

Aspect ratio

Resolution

Style LoRAs1/3

CINEMATIC FILMSTILL REDMOND FLUXKLEIN9B

Seed

Safety checkerOff

SpicyAPI does not run a safety checker on this model. Your prompt and media go to the model as written, and the model itself decides what comes back.

Checking your account before enabling submission and pricing.

Output

Ready for a new run

Example result

Animate

Edit

Split into layers

{
  "model": "black-forest-labs/flux-2-klein-9b-lora/text-to-image",
  "input": {
    "aspect_ratio": "16:9",
    "resolution": "2k",
    "loras": [
      {
        "path": "https://huggingface.co/artificialguybr/CINEMATIC-FILMSTILL-REDMOND-FLUXKLEIN9B/resolve/main/%5BFLUX.2.Klein%5DFilmStill_Redmond.safetensors",
        "scale": 0.9
      }
    ],
    "seed": 202,
    "prompt": "Cinematic, Film Still. Dusk on the top level of an empty multi-storey car park built of raw board-marked concrete, sodium lamps flickering on, a violet sky beyond the open edge. A woman in her early thirties with South Asian features and long dark hair leans against a concrete pillar in a black satin slip dress, one thin strap fallen to her upper arm, the wind lifting her hair, giving the camera a smouldering glance over her shoulder. Anamorphic 40mm lens, shallow depth of field, teal and amber grade, fine film grain."
  }
}

Introduction

FLUX.2 Klein 9B LoRA Text to image draws one image from a written prompt on Black Forest Labs' FLUX.2 [klein] 9B, the step-distilled 9B model, with up to three of your own LoRA files loaded from direct links. Choose one of nine aspect ratios at 1k or 2k and get a JPEG back, most often in about ten seconds of model time. The platform adds no content filter to the prompt, the weights or the image, and the price per image is flat.

Choose a mode by starting material

What you have	Mode
a written description and a LoRA	FLUX.2 Klein 9B LoRA Text to image
one to three images to change	FLUX.2 Klein 9B LoRA Image editing

Key features & highlights

Your LoRAs, both klein 9B builds

LoRAs trained on the distilled klein 9B and on klein Base 9B both loaded here in our tests. A LoRA trained on Base for the stronger levels changed what the model drew; a film-still LoRA shifted the grade of a whole scene toward teal and amber. FLUX.1 and klein 4B LoRAs, and files saved in diffusers PEFT format, ran without an error and changed nothing, so check the base model on the LoRA's card before you use it.

Fast enough to iterate

Most images took about ten seconds of model time, at 1k and at 2k alike, and 2k costs the same as 1k. That makes it practical to try a LoRA at three strengths and a few seeds in one sitting, then keep the best frame.

No content filter

Prompts and LoRA files are not reviewed. On its own the model holds back at the Strong and Very strong levels; LoRAs trained for that content on klein 9B rendered both in our tests. The Maximum level rendered without any LoRA.

Request parameters

Request parameters for FLUX.2 Klein 9B LoRA · Text to Image
Parameter	Required	Description
`prompt`	Yes	Required, up to 5,000 characters. Put the LoRA's trigger phrase first, then describe subject, clothing, place, light and lens in sentences. at least 1 characters · at most 5000 characters
`aspect_ratio`	No	Nine shapes from 21:9 to 9:21; 1:1 by default. The shape does not change the price. `1:1` · `16:9` · `9:16` · `4:3` · `3:4` · `3:2` · `2:3` · `21:9` · `9:21` · Default: `1:1`
`resolution`	No	1k (default) is about one megapixel, for example 1344 x 768 or 1024 x 1024; 2k puts 2048 pixels on the long side, for example 2048 x 1152. Both cost the same. `1k` · `2k` · Default: `1k`
`loras`	No	Up to three {path, scale} entries trained for FLUX.2 [klein] 9B, distilled or Base; scale 0 to 4, 1 by default. In our tests 0.5 was faint and 4 broke the picture. Default: `[]` at most 3 items
`seed`	No	Optional, 0 to 2,147,483,647. Reuse it to compare one change at a time; in our tests the same seed gave the same or a near-identical picture. `≥ 0` · `≤ 2147483647`

Pricing

$0.018 / image

Billed per image at one flat price. The aspect ratio, 1k or 2k and loading LoRAs do not change the price.

Use cases

Run a hub LoRA as an API call

Paste a direct .safetensors link from a model hub, set the scale and write the prompt. There is nothing to train, host or keep warm, and up to three LoRAs stack in one request.

Graded stills and styled series

A film or analog LoRA keeps a whole set of images in one grade across nine shapes, from 21:9 banners to 9:16 phone screens, at one flat price per image.

Character work with stacked LoRAs

Load a LoRA for the stronger levels trained for klein 9B with a character or style LoRA beside it, and keep every person an adult.

Explore key use-case tutorials

Prompt guide & pro tips

Trigger first, then sentences

Start with the LoRA's trigger phrase from its model card, then describe subject, clothing, place, light and lens in full sentences. The model follows long, specific prompts well, and there is no negative prompt, so say what you want rather than what to avoid.

Move the scale in small steps

Start at 0.8 to 1. In our tests 0.5 gave a faint version of a style and 4 broke the picture into blocks. Raise or lower by 0.2 and look at faces and hands first; they are the first things to distort.

Check that the LoRA is doing the work

Run the same prompt and seed with and without the file. Style LoRAs often refine a look the words already ask for, such as grain, grade or a pixel grid; LoRAs that teach new content change the picture outright. No visible change usually means the wrong base model.

Describe the look, light and lens

Without a LoRA the model makes its own wardrobe choices. Spend the prompt on the setting, the light and the lens, and keep the outfit description short and plain.

API quick start

# Generate the idempotency key once; reuse it when retrying after a lost response.
export SPICY_IDEMPOTENCY_KEY="${SPICY_IDEMPOTENCY_KEY:-$(uuidgen)}"
curl -X POST https://api.spicyapi.ai/api/v1/jobs/createTask \
  -H "Authorization: Bearer $SPICY_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $SPICY_IDEMPOTENCY_KEY" \
  -d '{
  "model": "black-forest-labs/flux-2-klein-9b-lora/text-to-image",
  "input": {
    "aspect_ratio": "1:1",
    "prompt": "A ceramic teapot on a linen cloth, soft window light, muted palette",
    "resolution": "1k"
  }
}'

# Copy data.taskId from the response into TASK_ID, then poll the status.
TASK_ID="TASK_ID_FROM_RESPONSE"
curl "https://api.spicyapi.ai/api/v1/jobs/recordInfo?taskId=$TASK_ID" \
  -H "Authorization: Bearer $SPICY_API_KEY"

See the API reference for authentication, the complete schema, polling, and error handling.

Missing the model you need?

Tell us what to add and where you would use it. Requests go to the people who decide what gets onboarded next.

LORA IMAGE API

FLUX.2 Klein 9B LoRA API at a glance

FLUX.2 [klein] 9B with up to three of your own LoRAs: fast text to image and editing at 1k or 2k, with no content filter.

LoRAs per call: 3; Up to three weight files per image, each with its own strength from 0 to 4.
Typical render: ~10 s; Median model time across our test runs at both size tiers, text to image and editing.
Largest size tier: 2k; 2048 pixels on the long side; 1k is about one megapixel. Both cost the same.
Edit references: 1 – 3; Combine or restyle up to three source images in one edit.
Klein 9B builds: Both; LoRAs trained on the distilled 9B or on Base 9B both loaded in our tests.
Content filter: None; No review of your LoRAs and no filter on the result.

Results

What you can make with FLUX.2 Klein 9B LoRA

Film look LoRA
A trigger phrase and a whole grade
A cinematic film-still LoRA at 0.9 with its trigger first: the empty car park, sodium lamps and violet sky come back with an anamorphic look and teal-and-amber grade, at 2048 by 1152.
Text to image with a film-still LoRA at 0.9, 16:9 at 2k.
Style LoRA
The same model, drawn in pixels
A pixel-art LoRA trained for klein 9B, with its trigger words, on a moonlit concrete amphitheatre. The words alone already gave pixel art; the LoRA coarsened the pixel grid and added dithered shading.
Text to image with a pixel-art LoRA at 0.8, 16:9.
Edit with two sources
Two portraits, one night out
Two separate head-and-shoulders photos became one couple at a lit fountain, with new clothes and an analog flash look from a LoRA.
Edit combining two portraits with an analog-photo LoRA, 3:2.

More clips made with FLUX.2 Klein 9B LoRA

Edit of one street photo with a film-still LoRA, auto shape.
Edit of one park photo without a LoRA, 16:9 at 2k.
Text to image with an analog-photo LoRA at 0.9, 2:3.
Text to image without a LoRA, 3:2 at 2k.

Content test

Is FLUX.2 Klein 9B LoRA uncensored? What we tested

We ran each level on our live setup and checked every clip frame by frame. A pass means the content was actually on screen, not just that the request went through.

Content levels tested on FLUX.2 Klein 9B LoRA, by mode
Level	Image editing	Text to image
MildThe lightest prompts in the test	RendersTested Oct 9, 2026	RendersTested Oct 9, 2026
ModerateWhere filtered models start to refuse	RendersTested Oct 9, 2026	RendersTested Oct 9, 2026
StrongThe middle of the scale	RendersTested Oct 9, 2026	RendersTested Oct 9, 2026
Very strongWhere most models start to hold back	RendersTested Oct 9, 2026	RendersTested Oct 9, 2026
MaximumThe top of the scale	RendersTested Oct 9, 2026	RendersTested Oct 9, 2026

Renders: on screen as written. Softened: generated, but milder than what was written. Refused: rejected. Not tested yet: we have not run that level on that mode.

How each mode behaves

All modes: Use LoRAs trained for FLUX.2 [klein] 9B, the distilled or the Base build. FLUX.1 [dev] and klein 4B LoRAs loaded without an error and changed nothing in our tests, and so did a file saved in diffusers PEFT format. A link that needs a login or points to a web page fails the request. Up to three LoRAs, scale 0 to 4. There is no negative prompt and no step or guidance control; output is JPEG, one image per request.
All modes: There is no content filter. On its own the model holds back: in our tests it toned the Strong level down by cropping or covering, softened a male subject the same way, and stopped short of the Very strong level. With LoRAs trained for that content on klein 9B loaded, both the Strong and Very strong levels rendered as asked. A heavy LoRA can also change a face, skin tone or framing, so lower the scale if the person drifts.
Image editing: Editing takes one to three sources. With auto the result keeps the first image's shape at about one megapixel; choose a ratio for the 1k or 2k tier. With two or more sources, faces and hair can drift, so check each person.
All modes: Every person you describe or upload must be an adult. You are responsible for the lawfulness of what you generate and publish, including the rights of anyone shown in a source image. Generated media is retained for 14 days and then deleted.

Use cases

What to build with the FLUX.2 Klein 9B LoRA API

The LoRA you already found
A klein 9B style, character or look from a model hub becomes a plain API call: paste the file link, set the scale, write the prompt.
Restyle and recast photos
Editing keeps a person and changes the outfit, the place or the light, with a LoRA setting the look.
Fast series and drafts
About ten seconds per image makes it practical to try many seeds, wordings and LoRA strengths in one session.
LoRAs for what the base holds back
Nothing is reviewed or filtered. LoRAs trained for klein 9B are how you get past what the base model holds back on its own.

Prompts

FLUX.2 Klein 9B LoRA prompt examples you can copy

Paste one into the playground as it is, then change one thing at a time.

Text to image
Rooftop car park at dusk
Cinematic, Film Still. Dusk on the top level of an empty multi-storey car park built of raw board-marked concrete, sodium lamps flickering on, a violet sky beyond the open edge. A woman in her early thirties with South Asian features and long dark hair leans against a concrete pillar in a black satin slip dress, one thin strap fallen to her upper arm, the wind lifting her hair, giving the camera a smouldering glance over her shoulder. Anamorphic 40mm lens, shallow depth of field, teal and amber grade, fine film grain.
Why it works. A film-still LoRA for klein 9B at 0.9 with its trigger "Cinematic, Film Still." first, 16:9 at 2k. Against the same seed without it, the LoRA pulled the grade toward teal and amber and deepened the shadows.
Try this prompt
Image editing
Trench coat to satin on the stairwell
Cinematic, Film Still. Keep her face, hair and expression. Move her to the landing of a raw concrete stairwell lit by a single hard shaft of light from above, and change the trench coat to a black satin slip dress with a low neckline. Deep shadows, teal and amber grade, film grain.
Why it works. One street photo, auto shape, the film-still LoRA at 0.8. Her face, hair and expression stay; the coat becomes a black satin slip dress and the street a concrete stairwell under one shaft of light.
This mode starts from your own image: add it in the playground before you run.
Try this prompt
Text to image
A shaft of light in a concrete stairwell
AnalogRedmAF, Analog. Looking up a raw concrete spiral stairwell, one hard shaft of daylight falling from a skylight. A man in his late thirties with dark brown skin and a close-cropped beard leans on the curved concrete balustrade in a cream linen shirt open at the collar, sleeves rolled to the elbow, looking down into the lens with a slow, confident half-smile. 35mm film photograph, warm grain, soft halation, deep shadows.
Why it works. An analog-photo LoRA at 0.9 with its trigger first, 2:3 at 1k. It added warm grain and halation; the same seed without the LoRA came out cleaner and cooler.
Try this prompt

Pricing

FLUX.2 Klein 9B LoRA API pricing

One image starts at $0.018.

The playground shows the exact price for your settings before you run. Failed or timed-out tasks are usually not charged. Some models charge for content refusals as stated on their model page; check those billing terms before submitting.

See every price tier

Get started

How to use the FLUX.2 Klein 9B LoRA API in three steps

Step 1
Pick LoRAs made for klein 9B
Look for files whose base model is FLUX.2 [klein] 9B or FLUX.2 [klein] Base 9B; in our tests both kinds loaded. FLUX.1 LoRAs and klein 4B LoRAs ran without an error and changed nothing.
Step 2
Trigger phrase first
Most klein LoRAs name a trigger on their model card, such as "Cinematic, Film Still." Start the prompt with it, then describe subject, clothing, place and light in sentences.
Step 3
Start the scale at 0.8 to 1
In our tests 0.5 gave a faint version of a style, 1 gave the full look, and 4 broke the picture into blocks. Move in steps of 0.2 and stop where faces still hold.

Open the playground Get an API key

Parameter reference and code samples are in the API and Readme tabs at the top of this page.

Why SpicyAPI

Why run FLUX.2 Klein 9B LoRA on SpicyAPI

No filter from us
We add no content filter of our own. What the model renders is what you get back.
Tested, not guessed
Content labels come from real generations we ran and checked, with the test date on the page.
Pay per result, in dollars
No subscription and no credits. Failed or timed-out tasks are usually not charged. Some models charge for content refusals as stated on their model page; check those billing terms before submitting.
One key for every model
One API key and balance cover image, video, audio and chat models, so you can explore more models from the same account.

FAQ

FLUX.2 Klein 9B LoRA API FAQ

About FLUX.2 Klein 9B LoRA

FLUX.2 Klein 9B LoRA puts the fast, step-distilled klein 9B behind your own LoRAs: text to image and editing at 1k or 2k, three LoRAs per call, about ten seconds per image and no content filter. Bring LoRAs trained for klein 9B, distilled or Base.