Flux General
FLUX.1 Dev with unified LoRA, ControlNet, IP-Adapter, and reference guidance plus inpainting and inversion endpoints.
Studio is in early access; availability resolves per project after sign-in.
- Endpoints
- 5
- Eligible
- 1
- Input schemas imported
- 5 / 5
- Delivery
- fal.ai
- Weights
- Not stated
- Developer
- Not stated in source
Overview
Flux General wraps FLUX.1 [dev] with LoRA, ControlNet, IP-Adapter, and reference-only guidance in one API for fine-grained creative direction. Classical 3.5 CFG applies with the XLabs IP-Adapter; inference defaults to 28 steps over a 1-50 range with 1-10 image batches. Inpainting, differential diffusion, image-to-image, and RF-inversion endpoints extend coverage.
Capabilities
- Four-system unified control API
- 28-step default over 1-50 range
- 1-10 image batch generation
- Inpainting, inversion, and i2i endpoints
Best for
- Brand-consistent product visualization
- Pose-controlled character generation
- Reference-guided style transfer
Use cases
- Single-endpoint directed shoots
- Production control workflows
- Multi-variant batch runs
Endpoints
4 under review · 1 eligible. Cataloged is not the same as executable: Studio resolves which endpoints a project can run.
| Endpoint | Task | Catalog status | Inputs |
|---|---|---|---|
fal-ai/flux-general | unknown | Under review | 33 fields · requires prompt |
fal-ai/flux-general/differential-diffusion | unknown | Under review | 36 fields · requires prompt, image_url, change_map_image_url |
fal-ai/flux-general/image-to-image | image to image | Eligible | 35 fields · requires prompt, image_url |
fal-ai/flux-general/inpainting | unknown | Under review | 36 fields · requires prompt, image_url, mask_url |
fal-ai/flux-general/rf-inversion | unknown | Under review | 36 fields · requires prompt, image_url |
Questions
Which control systems combine?
LoRA, ControlNet, IP-Adapter, and reference-only guidance.
Beyond the base endpoint?
Differential diffusion, image-to-image, inpainting, and RF inversion.