Which LoRAs work here?

LoRAs trained for FLUX 3 Klein, including those trained on its Base build: in our tests a pastel style LoRA trained on Base changed the result here at every strength we tried. FLUX.2 klein LoRAs ran without an error and changed nothing.

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 already rendered NSFW portraits of adult women, but the strongest requests came back toned down; LoRAs trained for unfiltered work on FLUX 3 Klein are the way past those limits.

Why did my LoRA request fail?

Usually the link points to a repository page instead of the file, or the file cannot be downloaded. Use a public URL that downloads the .safetensors file itself. Failed requests are not charged.

My LoRA loaded but nothing changed. Why?

In our tests that happened with LoRAs trained for another model (FLUX.2 klein) and with two LoRAs saved in diffusers format. Neither raises an error. Compare the same seed with and without the file, and try a higher scale in editing.

Per image, by size tier. In editing, each reference image after the first adds a small charge. The number of LoRAs, the aspect ratio and the file format do not change the price; failed requests are not charged.

FLUX 3 Klein LoRA API — Uncensored NSFW Images, Your LoRAs

black-forest-labs/flux-3-klein-lora/edit

FLUX 3 Klein LoRA image editing changes one to three reference images on FLUX 3 Klein with up to three of your own LoRA files. Name each source in the prompt and let a LoRA set the look, with no content filter.

From $0.025/image

Run model View API

Input

Image editing

Prompt289/5000

Reordering updates prompt references. Removing media also removes its references; review your prompt.

Imagesaffects price2/3

Drop files or media links, paste, or browse · Up to 3 files · Max 10 MB · JPG · PNG · WEBP · GIF

<ref_image_1>
<ref_image_2>

Aspect ratio

Resolutionaffects price

Style LoRAs1/3

civitai · 3409162

Output format

JPEGPNGWebP

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-3-klein-lora/edit",
  "input": {
    "image_urls": [
      "https://cdn.spicyapi.ai/models/examples/inputs/7c528e77c8f66555.jpeg",
      "https://cdn.spicyapi.ai/models/examples/inputs/640778e4e9fe0b48.jpeg"
    ],
    "aspect_ratio": "auto",
    "resolution": "1k",
    "loras": [
      {
        "path": "https://civitai.com/api/download/models/3409162",
        "scale": 1.2
      }
    ],
    "output_format": "jpeg",
    "seed": 101,
    "prompt": "pastelstyle, Put the straw panama hat from <ref_image_2> on the man from <ref_image_1>, tilted low over one eye. Keep his face and curls. Open his linen jacket and turn the market into a soft pastel watercolour scene in early evening light; he gives the viewer a slow, charming half smile."
  }
}

Introduction

FLUX 3 Klein LoRA Image editing changes one to three reference images you upload on Black Forest Labs' FLUX 3 Klein, the 5.6B open-weights model released in October 2026, in its step-distilled build, with up to three of your own LoRA files loaded from direct links. Choose 1k or 2k and one of fifteen aspect ratios, get a JPEG, PNG or WebP back, and pay per image. The platform adds no content filter.

Choose a mode by starting material

What you have	Mode
a written description	FLUX 3 Klein LoRA Text to image
one to three images to change	FLUX 3 Klein LoRA Image editing

Key features & highlights

Your LoRAs on the newest Klein

A pastel style LoRA trained on FLUX 3 Klein Base loaded on this distilled build in our tests and changed the look at every strength: 0.5 a light wash, 1 a full illustration, 2 a heavy cartoon style. FLUX.2 klein LoRAs ran without an error and changed nothing, and two LoRAs saved in diffusers format changed almost nothing, so compare the same seed with and without the file.

Name each source

Write <ref_image_1>, <ref_image_2> and <ref_image_3> in the prompt, in upload order, the notation Black Forest Labs documents for this model. In our tests a garment or a hat from one photo went onto a person from another, with the face kept.

No content filter

Prompts, sources, LoRA files and results are not reviewed. In our tests unfiltered portraits of adult women came through and unfiltered edits were carried out as asked; only the strongest requests came back toned down by the model itself, and a LoRA trained for unfiltered work on FLUX 3 Klein is the way past that.

Request parameters

Request parameters for FLUX 3 Klein LoRA · Image Editing
Parameter	Required	Description
`prompt`	Yes	Required, up to 5,000 characters. Put the LoRA's trigger word first, name sources as <`ref_image_1`>, <`ref_image_2`>, <`ref_image_3`>, and say what must stay the same. at least 1 characters · at most 5000 characters
`image_urls`	Yes	Required, one to three public image links, up to 10 MB each, in the order your prompt refers to them. A link that does not open or is not an image fails the request. at least 1 items · at most 3 items
`aspect_ratio`	No	auto (default) keeps the first image's shape; or pick one of fifteen shapes from 3:1 to 1:3. Does not change the price. `auto` · `1:1` · `16:9` · `9:16` · `4:3` · `3:4` · `3:2` · `2:3` · `4:5` · `5:4` · `21:9` · `9:21` · `2:1` · `1:2` · `3:1` · `1:3` · Default: `auto`
`resolution`	No	1k (default) is about one megapixel; 2k is about four. 2k costs more. `1k` · `2k` · Default: `1k`
`loras`	No	Up to three {path, scale} entries trained for FLUX 3 Klein or its Base build; scale 0 to 4. Default: `[]` at most 3 items
`output_format`	No	jpeg (default), png or webp. Same price for all three. `jpeg` · `png` · `webp` · Default: `jpeg`
`seed`	No	Optional, 0 to 2,147,483,647. Reuse it to compare LoRA strengths on the same sources. `≥ 0` · `≤ 2147483647`

Pricing

Billed per image by size tier (1k or 2k); each reference image after the first adds a small charge. Aspect ratio, file format and the number of LoRAs do not change the price.

Pricing
Specification	Price
`1k`	$0.0252 / image
`2k`	$0.0828 / image

Use cases

Outfit and prop swaps

Put a dress, hat or jacket from a product photo on a person from another photo in one request.

Relighting and restyling

Turn morning into sunset or noon into blue hour while the face and the room stay.

Hub LoRAs as an API

Paste a direct link to a FLUX 3 Klein LoRA, set the strength and write the prompt; up to three LoRAs stack in one request.

Explore key use-case tutorials

Prompt guide & pro tips

Write a full caption

Start with the LoRA's trigger word, then describe framing, subject, clothing, place, light and lens in full sentences. There is no negative prompt, so say what you want rather than what to avoid.

List what stays

Next to the change, name what must not move: face, hair, pose, the room. Without it the model is free to reinterpret the whole frame.

Move the strength in small steps

Start at 0.6 to 1 in text to image and 1.2 to 1.5 in editing, and watch faces first: a strong style LoRA can make an adult face look cartoonish.

Draft at 1k, finish at 2k

Fix the seed while you work at 1k, then rerun the same inputs at 2k for the final image.

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-3-klein-lora/edit",
  "input": {
    "aspect_ratio": "auto",
    "image_urls": [
      "https://example.com/reference.png"
    ],
    "output_format": "jpeg",
    "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 3 Klein LoRA API at a glance

FLUX 3 Klein with up to three of your own LoRAs: text to image and multi-reference 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.
LoRAs that load: Base; A LoRA trained on FLUX 3 Klein Base changed the result here in our tests.
Aspect ratios: 15; From 3:1 to 1:3, every one at 1k or 2k.
Edit references: 1 – 3; Name each source in the prompt as <ref_image_1>, <ref_image_2>, <ref_image_3>.
Largest tier: ~4 MP; 2k is about four megapixels; 1k is about one.
Content filter: None; No review of your LoRAs and no filter on the result.

Results

What you can make with FLUX 3 Klein LoRA

Two references and a LoRA
A hat from one photo, a pastel finish
The straw hat from <ref_image_2> went onto the man in <ref_image_1>, and a pastel style LoRA at 1.2 washed the market into soft watercolour tones.
Edit with two references and a pastel style LoRA at 1.2.
Style LoRA
A light pastel wash at 0.5
The same pastel LoRA at half strength keeps a photograph a photograph and only lifts the palette toward lilac and cream.
Text to image with a pastel style LoRA at 0.5, 3:4.
Edit with a LoRA
A photo turned illustration
One photo of a woman reading on a stone threshing floor, restyled as a pastel illustration with a new off-the-shoulder dress.
Edit of one photo with a pastel style LoRA at 1.2.

More clips made with FLUX 3 Klein LoRA

Text to image with a pastel style LoRA at 0.6, 16:9.
Edit of one photo without a LoRA, auto shape.
Text to image without a LoRA, 4:5.

Content test

Is FLUX 3 Klein 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 3 Klein LoRA, by mode
Level	Image editing	Text to image
MildThe lightest prompts in the test	RendersTested Oct 16, 2026	RendersTested Oct 16, 2026
ModerateWhere filtered models start to refuse	RendersTested Oct 16, 2026	RendersTested Oct 16, 2026
StrongThe middle of the scale	RendersTested Oct 16, 2026	RendersTested Oct 16, 2026
Very strongWhere most models start to hold back	Not tested yet	Not tested yet
MaximumThe top of the scale	Not tested yet	Not tested yet

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 3 Klein or FLUX 3 Klein Base. FLUX.2 klein LoRAs loaded without an error and changed nothing in our tests, and two diffusers-format LoRAs changed almost nothing. A link to a repository page or a file that cannot be downloaded fails the request. Up to three LoRAs, scale 0 to 4. No negative prompt, no step or guidance control; one image per request.
All modes: There is no content filter, but the model has limits of its own. Without a LoRA, unfiltered portraits of women came through with fine detail softened, while the strongest requests came back toned down. A strong style LoRA can also soften what the prompt asks for.
Image editing: Editing takes one to three sources. With auto the result keeps the shape of the first image. Style LoRAs act more weakly in editing than in text to image; we needed about 1.2 to 1.5 for a clear change.
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 3 Klein LoRA API

Run a hub LoRA as an API call
Paste a direct link to a FLUX 3 Klein LoRA, set the strength and write the prompt. Nothing to train, host or keep warm.
Restyle photos with a look
Editing keeps a person and changes the outfit or the light while a style LoRA sets the palette.
Combine sources and a style
Put a hat from one photo on a man from another and render the result in your LoRA's style in one request.
Your own LoRAs, unfiltered
LoRA files and results are not reviewed. Bring unfiltered LoRAs trained for FLUX 3 Klein and keep every person an adult.

Prompts

FLUX 3 Klein LoRA prompt examples you can copy

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

Image editing
Hat from one photo, style from a LoRA
pastelstyle, Put the straw panama hat from <ref_image_2> on the man from <ref_image_1>, tilted low over one eye. Keep his face and curls. Open his linen jacket and turn the market into a soft pastel watercolour scene in early evening light; he gives the viewer a slow, charming half smile.
Why it works. Trigger word first, then the two sources named <ref_image_1> and <ref_image_2>. A pastel LoRA at 1.2; in editing it needed more strength than in text to image.
This mode starts from your own image: add it in the playground before you run.
Try this prompt
Text to image
Village well in a pastel wash
pastelstyle, photograph of a woman in her late thirties with fine laugh lines and long wavy auburn hair, sitting on the rim of an old stone well in a sun-bleached village square, wearing a pale lilac linen sundress with one strap slipped from her shoulder, almond blossom drifting around her, half-lidded alluring glance, soft pastel palette, airy morning light.
Why it works. The pastel LoRA at 0.5 with its trigger first: still a photograph, with the palette pulled toward lilac and cream.
Try this prompt
Text to image
Figs under the pergola
Cinematic photograph under a vine-covered pergola at dusk on a Greek island. An adult couple in their thirties, a Greek woman in a rust-red wrap dress and a Moroccan man in an open cream shirt, share a ripe fig across the corner of a long rustic table, her fingers at his lips, both laughing softly. Strings of warm bulbs, terracotta jugs, indigo sky, 50mm lens, film grain.
Why it works. No LoRA: the distilled model on its own, two people, warm bulbs against an indigo sky.
Try this prompt

Pricing

FLUX 3 Klein LoRA API pricing

One image starts at $0.018 at 1K and $0.054 at 2K.

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 3 Klein LoRA API in three steps

Step 1
Pick LoRAs made for FLUX 3 Klein
Look for a base model of FLUX 3 Klein or FLUX 3 Klein Base. FLUX.2 klein LoRAs ran without an error here and changed nothing. The model is new, so the choice of LoRAs is still small.
Step 2
Trigger word first
Start the prompt with the trigger word from the LoRA's card, such as "pastelstyle", then describe the scene in full sentences.
Step 3
Find the strength by eye
With a pastel style LoRA, 0.5 gave a light wash, 1 a full illustration and 2 a heavy, cartoon-like look. Start at 0.6 to 1 for photos and stop where faces still read as adults.

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 3 Klein 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 3 Klein LoRA API FAQ

About FLUX 3 Klein LoRA

FLUX 3 Klein LoRA puts Black Forest Labs' newest open-weights image model behind your own LoRAs: text to image and multi-reference editing at 1k or 2k, three LoRAs per call and no content filter.