# PhotoMaker - Core.Today AI API

> Generate stylized photos of a person from 1-4 reference photos (9M+ runs). Ten styles including Cinematic, Disney Character, and Digital Art — keep the identity, change everything else.

- **Provider**: Tencent ARC
- **Model ID**: tencentarc/photomaker
- **Category**: Image Generation
- **Credits**: 10 per image (num_outputs billed per image)
- **Speed**: Medium
- **Quality**: High

## Features
- Identity-preserving generation from 1-4 photos
- 10 built-in styles
- Prompt-controlled scenes and outfits
- Up to 4 outputs per run (billed per image)

## Use Cases
- Profile avatar packs
- Team photo stylization
- Marketing persona imagery
- Yearbook/event themed portraits

## API Endpoint

Base URL: https://api.core.today/v1
Create Prediction: POST /predictions
Get Status: GET /predictions/{job_id}

## Authentication
Header: X-API-Key: YOUR_API_KEY

## Input Parameters

### Required
- **input_image**: string - Reference photo of the person (input_image2-4 for more)

### Optional
- **prompt**: string (default: A photo of a person img) - Scene description. Must include the trigger word 'img' after the class noun (e.g. 'a photo of a man img')
- **input_image2**: string - Additional input image (optional)
- **input_image3**: string - Additional input image (optional)
- **input_image4**: string - Additional input image (optional)
- **style_name**: string (default: Photographic (Default)) - Built-in style preset
  Options: (No style), Cinematic, Disney Charactor, Digital Art, Photographic (Default), Fantasy art, Neonpunk, Enhance, Comic book, Lowpoly, Line art
- **negative_prompt**: string (default: nsfw, lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry) - Negative Prompt. The negative prompt should NOT contain the trigger word.
- **num_steps**: integer (default: 20) - Denoising steps (1-100)
- **style_strength_ratio**: number (default: 20) - Style strength (15-50). Higher applies the style harder
- **num_outputs**: integer (default: 1) - Number of images (1-4). Billed per image
- **guidance_scale**: number (default: 5) - Guidance scale. A guidance scale of 1 corresponds to doing no classifier free guidance.
- **seed**: integer - Random seed
- **disable_safety_checker**: boolean (default: false) - Disable safety checker for generated images.

## Examples

### Cinematic Avatar
Cinematic-style portraits from one reference photo
```json
{
  "model": "tencentarc/photomaker",
  "input": {
    "input_image": "https://example.com/selfie.jpg",
    "prompt": "portrait photo of a man img, city night background, dramatic lighting",
    "style_name": "Cinematic",
    "num_outputs": 2
  }
}
```

## Response Format
```json
{
  "job_id": "abc123",
  "status": "pending | processing | completed | failed",
  "result": "URL or data (when completed)"
}
```

## Usage Flow
1. POST /predictions with model and input -> receive job_id
2. GET /predictions/{job_id} -> poll until status is completed or failed
3. Result contains output URL(s)

## Tips
- The prompt MUST contain the trigger word 'img' right after the person noun — 'a man img'
- 2-4 reference photos of the same person noticeably improve identity fidelity
- Raise style_strength_ratio toward 40 for bolder stylization
- Only generate images of people who have consented

## Documentation
https://replicate.com/tencentarc/photomaker