Nai-Diffusion-4-5-Curated - Text To Image - gptproto

Authentication

Sign up for a GPTProto account at https://gptproto.com
Navigate to the API Keys section in your dashboard
Generate a new API key (sk-xxxxx)
Copy and securely store your API key For authentication details, please refer to the Authentication section.

Initiate Request

curl --location 'https://gptproto.com/api/v3/novel/nai-diffusion-4-5-curated/text-to-image' \
--header 'Authorization: GPTPROTO_API_KEY' \
--header 'Content-Type: application/json' \
--data '{
  "prompt": "1girl, cute, anime style, blue eyes, smile",
  "negative_prompt": "lowres, bad anatomy, text, error",
  "size": "1400x1400",
  "n": 1,
  "quality_toggle": true,
  "steps": 23,
  "scale": 5,
  "sampler": "k_euler_ancestral"
}'

Query Result

If the request that generated your content includes the parameter enable_sync_mode set to true (some models do not support this parameter, but you still need to query the result by id), you must call the Query Result endpoint to retrieve the final output.

curl -X GET "https://gptproto.com/api/v3/predictions/{id}/result" \
  -H "Authorization: GPTPROTO_API_KEY" \
  -H "Content-Type: application/json"

After submitting your prediction request, the response will contain the id you need.
You can find it in either:

data.id – the unique identifier of the prediction
data.urls[0].get – a ready-to-use GET URL that already embeds the id

Response Example

{
    "data": {
        "id": "abc",
        "model": "model_name",
        "outputs": [],
        "urls": {
            "get": "https://gptproto.com/api/v3/predictions/abc/result"
        },
        "status": "completed",
        "error": null,
        "executionTime": 0,
        "timings": {
            "inference": 0
        },
        "has_nsfw_contents": [],
        "created_at": "2026-01-01 00:00:00"
    },
    "message": "success",
    "code": 200
}

Parameters

Request Body

Core Parameters

Parameter	Type	Required	Default	Range	Description
`prompt`	string	✅ Yes	-	-	Image description prompt
`action`	string	❌ No	`generate`	`generate` `img2img` `infill`	Generation mode
`size`	string	❌ No	`832x1216`	`64x64` - `49152x49152`	Image size in format `widthxheight`. See Common Sizes for recommended values
`n`	integer	❌ No	`1`	`1` - `4`	Number of images to generate

Prompt Control Parameters

Parameter	Type	Required	Default	Range	Description
`negative_prompt`	string	❌ No	`""`	-	Negative prompt to avoid certain elements
`quality_toggle`	boolean	❌ No	`true`	-	Automatically add quality tags
`uc_preset`	integer	❌ No	`0`	`0` - `3`	Negative content preset strength: `0` = Heavy, `1` = Light, `2` = Human Focus, `3` = None

AI Generation Control Parameters

Parameter	Type	Required	Default	Range	Description
`steps`	integer	❌ No	`28`	`1` - `50`	Sampling steps
`scale`	float	❌ No	`6.0`	`0` - `10`	CFG Scale (prompt guidance strength)
`sampler`	string	❌ No	`k_euler_ancestral`	See Samplers	Sampling method
`noise_schedule`	string	❌ No	`karras`	`native` `karras` `exponential`	Noise schedule algorithm
`seed`	integer	❌ No	Random	`0` - `4294967295`	Random seed for reproducibility
`cfg_rescale`	float	❌ No	`0`	`0` - `1`	CFG rescale value

Advanced Parameters

Parameter	Type	Required	Default	Description
`dynamic_thresholding`	boolean	❌ No	`false`	Enable dynamic thresholding (Decrisper)
`auto_smea`	boolean	❌ No	`false`	Automatically enable SMEA (V4/V4.5)
`skip_cfg_above_sigma`	integer	❌ No	`null`	Variety Boost (diversity enhancement)
`legacy`	boolean	❌ No	`false`	Use legacy mode
`use_coords`	boolean	❌ No	`false`	Use coordinate positioning (for multi-character)

Image-to-Image Parameters

These parameters are only used when action is set to img2img

{
  "prompt": "1girl, cute, anime style, colorful",
  "action": "img2img",
  "image": "https://example.com/image.jpg",
  "strength": 0.7,
  "size": "832x1216"
}

Parameter	Type	Required	Default	Range	Description
`image`	string	✅ Yes (when action=img2img)	-	-	Input image (Base64 encoded or URL)
`strength`	float	❌ No	`0.7`	`0.01` - `0.99`	Transformation strength. Higher values result in more changes
`noise`	float	❌ No	`0`	`0` - `0.99`	Additional noise for increased randomness

Vibe Transfer Parameters

{
  "prompt": "1girl, beautiful landscape, fantasy art",
  "size": "832x1216",
  "vibe_transfer": [
    {
      "image": "https://example.com/reference.jpg",
      "strength": 0.7,
      "information_extracted": 1.0
    }
  ]
}

Parameter	Type	Required	Default	Range	Description
`vibe_transfer`	array	❌ No	-	-	Array of style reference images
`vibe_transfer[].image`	string	✅ Yes	-	-	Reference image (Base64 or URL)
`vibe_transfer[].information_extracted`	float	❌ No	`1.0`	`0.01` - `1.0`	Information extraction level
`vibe_transfer[].strength`	float	❌ No	`0.6`	`0.01` - `1.0`	Style transfer strength

Multi-Character Generation (V4.5 Only)

Multi-character features are only supported by V4.5 models (nai-diffusion-4-5-full or nai-diffusion-4-5-curated)

{
  "prompt": "2girls, anime style, outdoor",
  "size": "1216x832",
  "character_prompts": [
    {
      "prompt": "1girl, long hair, blue eyes, smile",
      "uc": "lowres, bad anatomy",
      "center": {
        "x": 0.3,
        "y": 0.5
      },
      "enabled": true
    },
    {
      "prompt": "1girl, short hair, red eyes, serious",
      "uc": "lowres, bad anatomy",
      "center": {
        "x": 0.7,
        "y": 0.5
      },
      "enabled": true
    }
  ],
  "use_coords": true
}

Parameter	Type	Required	Default	Range	Description
`character_prompts`	array	❌ No	-	-	Array of character prompts
`character_prompts[].prompt`	string	✅ Yes	-	-	Character-specific prompt
`character_prompts[].uc`	string	❌ No	`"lowres, aliasing,"`	-	Character-specific negative prompt
`character_prompts[].center`	object	❌ No	-	-	Character position coordinates
`character_prompts[].center.x`	float	❌ No	`0.5`	`0.1` - `0.9`	X coordinate, `0.5` = horizontal center
`character_prompts[].center.y`	float	❌ No	`0.5`	`0.1` - `0.9`	Y coordinate, `0.5` = vertical center
`character_prompts[].enabled`	boolean	❌ No	`true`	-	Enable this character

ControlNet Parameters (V4/V4.5 Only)

ControlNet is only supported by V4 and V4.5 models

{
  "prompt": "beautiful landscape, detailed, high quality",
  "controlnet_model": "hed",
  "controlnet_condition": "https://example.com/control-image.jpg",
  "controlnet_strength": 1.2,
  "size": "1216x832"
}

Parameter	Type	Required	Default	Range	Description
`controlnet_model`	string	❌ No	-	See below	ControlNet model type
`controlnet_condition`	string	❌ No	-	-	Control condition image (Base64 or URL)
`controlnet_strength`	float	❌ No	`1.0`	`0.1` - `2.0`	Control strength

ControlNet Models:

hed - Palette Swap
midas - Form Lock (depth-based)
fake_scribble - Scribbler
mlsd - Building Control (line-based)
uniformer - Landscaper (segmentation-based)

Director Reference Parameters (V4/V4.5 Only)

Director Reference provides more precise character and style control. Supported models: nai-diffusion-4-5-full, nai-diffusion-4-5-curated, nai-diffusion-4-full, nai-diffusion-4-curated-preview

{
  "prompt": "solo, 1girl, cute, chibi, very aesthetic, masterpiece",
  "action": "img2img",
  "size": "896x1472",
  "image": "https://example.com/reference.jpg",
  "strength": 0.7,
  "steps": 23,
  "scale": 5.0,
  "sampler": "k_euler_ancestral",
  "director_reference_descriptions": [
    {
      "caption": {
        "base_caption": "character&style",
        "char_captions": []
      },
      "legacy_uc": false
    }
  ],
  "director_reference_information_extracted": [1],
  "director_reference_strength_values": [1],
  "director_reference_secondary_strength_values": [0]
}

Parameter	Type	Required	Range	Description
`director_reference_descriptions`	array	❌ No	-	Array of reference descriptions
`director_reference_descriptions[].caption`	object	✅ Yes	-	Caption object
`director_reference_descriptions[].caption.base_caption`	string	✅ Yes	-	Reference description (e.g., “character&style”, “character1”)
`director_reference_descriptions[].caption.char_captions`	array	❌ No	-	Character caption array, typically empty
`director_reference_descriptions[].legacy_uc`	boolean	❌ No	-	Use legacy negative prompt, default `false`
`director_reference_information_extracted`	array	❌ No	`0.0` - `1.0`	Array of information extraction levels
`director_reference_strength_values`	array	❌ No	`0.0` - `1.0`	Array of primary strength values
`director_reference_secondary_strength_values`	array	❌ No	`0.0` - `1.0`	Array of secondary strength values, typically `0`

Supported Samplers

Sampler	Description
`k_euler_ancestral`	Euler Ancestral (Default, Recommended)
`k_euler`	Euler
`k_dpmpp_2m`	DPM++ 2M
`k_dpmpp_sde`	DPM++ SDE
`k_dpmpp_2s_ancestral`	DPM++ 2S Ancestral
`ddim_v3`	DDIM

Common Image Sizes

Size Format	Aspect Ratio	Use Case
`832x1216`	Portrait (9:13)	Vertical character portraits (Default)
`1024x1024`	Square (1:1)	General purpose
`1216x832`	Landscape (13:9)	Horizontal landscapes
`1472x832`	Widescreen (16:9)	Widescreen backgrounds

Error Codes

Common Error Codes

Error Code	Error Name	Description
401	Unauthorized	API key is missing or invalid
403	Forbidden	Your API key doesn’t have permission to access this resource, or insufficient balance for the requested operation
429	Too Many Requests	You’ve exceeded your rate limit
500	Internal server error	An internal server error occurred
503	Content policy violation	Content blocked due to safety concerns (actual status code is 400)

​Authentication

​Initiate Request

​Query Result

​Parameters

​Request Body

​Core Parameters

​Prompt Control Parameters

​AI Generation Control Parameters

​Advanced Parameters

​Image-to-Image Parameters

​Vibe Transfer Parameters

​Multi-Character Generation (V4.5 Only)

​ControlNet Parameters (V4/V4.5 Only)

​Director Reference Parameters (V4/V4.5 Only)

​Supported Samplers

​Common Image Sizes

​Error Codes

​Common Error Codes

Authentication

Initiate Request

Query Result

Parameters

Request Body

Core Parameters

Prompt Control Parameters

AI Generation Control Parameters

Advanced Parameters

Image-to-Image Parameters

Vibe Transfer Parameters

Multi-Character Generation (V4.5 Only)

ControlNet Parameters (V4/V4.5 Only)

Director Reference Parameters (V4/V4.5 Only)

Supported Samplers

Common Image Sizes

Error Codes

Common Error Codes