Generate images from text prompts using Hugging Face's Inference API — directly from your LM Studio chat. Supports FLUX, SDXL, and other text-to-image models, plus LoRA adapters for style customization.
cd image npm install npx tsc
Load the built plugin in LM Studio.
| Field | Default | Description |
|---|---|---|
| HuggingFace API Token | (blank) | Required. Your HF access token from huggingface.co/settings/tokens. Use a token with at least read scope. |
| Default Model | black-forest-labs/FLUX.1-schnell | HuggingFace model ID for generation. Overridable per call. |
| Output Directory | ~/hf-images | Where generated images are saved. Created automatically. Supports ~/ prefix. |
.png or .jpeg fileFree tier note: HuggingFace free accounts can use FLUX.1-schnell and SDXL. Cold-start on an inactive model may take 20–60 seconds on the first call. Subsequent calls are fast.
Pro tier note: FLUX.1-dev and FLUX.2-dev require HuggingFace Pro credits. They produce noticeably higher quality output but cost credits per generation.
generate_image — Generate from textCreates an image from a text prompt and saves it to disk.
| Parameter | Default | Description |
|---|---|---|
prompt | (required) | Text description. Be specific — subject, style, lighting, mood, quality terms. |
model_id | (config default) | Override the default model for this call. |
negative_prompt | "" | What to exclude (e.g. "blurry, low quality, watermark"). Not all models support this. |
lora_id | "" | HuggingFace model ID of a LoRA adapter. FLUX models only. Use list_loras to find IDs. |
lora_scale | 1.0 | LoRA strength. 0.5–1.0 for subtle effects; 1.0–1.5 for strong. |
Returns:
file_path — absolute path of the saved imagemodel_used — which model was usedlora_used — LoRA adapter applied, if anyfile_size_bytes, mime_typeLoRA note: LoRA generation routes through the fal-ai provider, which supports FLUX LoRA compositing. Pair FLUX LoRAs with a FLUX base model.
list_models — Show available modelsReturns the curated list of popular text-to-image models with speed, style, and access tier.
Use when the user asks what models are available or wants to pick a different one.
Available models:
| Model ID | Speed | Access | Notes |
|---|---|---|---|
black-forest-labs/FLUX.1-schnell | Fast | Free | State-of-the-art quality, free tier friendly |
black-forest-labs/FLUX.1-dev | Medium | Pro | Higher quality than schnell, ~50 steps. Non-commercial. |
black-forest-labs/FLUX.2-klein | Fast (<1s) | Free | Real-time generation. Apache 2.0, fully open. |
black-forest-labs/FLUX.2-dev | Slow | Pro | 32B parameter flagship. Best quality. Non-commercial. |
stabilityai/stable-diffusion-xl-base-1.0 | Medium | Free | SDXL 1024px — stable and widely supported. |
You can also pass any HuggingFace text-to-image model ID directly to generate_image without it appearing in this list.
list_loras — Search LoRA adaptersSearches HuggingFace for LoRA adapters compatible with FLUX base models.
| Parameter | Default | Description |
|---|---|---|
search | "" | Keyword to filter (e.g. "anime", "portrait", "watercolor"). Blank returns popular FLUX LoRAs. |
Returns a list of LoRA model IDs sorted by downloads. Pass the id field as lora_id in generate_image.
Photorealistic portrait:
"A cinematic close-up portrait of a woman in golden hour light, soft bokeh background, 85mm lens, film grain"
Concept art:
"A futuristic megacity at night, neon reflections in rain puddles, cyberpunk aesthetic, ultra-detailed, 4K"
With negative prompt:
prompt:
"A peaceful mountain lake at sunrise"negative_prompt:"people, cars, buildings, text, watermark"
With LoRA style:
First:
list_loras("anime")→ find a LoRA ID Then:generate_image(prompt="...", lora_id="alvdansen/flux-koda", lora_scale=0.8)
Using a specific model:
generate_image(prompt="...", model_id="stabilityai/stable-diffusion-xl-base-1.0")
Some models require accepting their license on HuggingFace before they can be used:
huggingface.co/black-forest-labs/FLUX.1-schnell)read scopeThis only needs to be done once per model per account.
ultra-detailed, 8K, cinematic, professional photography, sharp focusin the style of Studio Ghibli, watercolor illustration, photorealistic, oil paintingblurry, low quality, deformed hands, text, watermark, oversaturatedgenerate_image(prompt, model_id?, negative_prompt?, lora_id?, lora_scale?)
list_models()
list_loras(search?)