# PixelAPI — AI Image Generation, Image Editing & Audio Generation API

> PixelAPI is a developer-friendly REST API offering 10 AI models for image generation, image editing, and audio generation. Built on self-hosted dedicated NVIDIA RTX GPUs, PixelAPI is significantly cheaper than cloud alternatives. Pay-per-use credit system: 1 credit = $0.001, most image operations cost just 3 credits ($0.003/image).

## Overview

PixelAPI provides 10 AI models accessible via a single API key. Unlike platforms like Replicate that run on rented cloud GPUs, PixelAPI runs on dedicated self-hosted hardware (NVIDIA RTX 4070 Ti Super 16GB, RTX 4060 Ti 16GB), eliminating cloud markup and keeping prices low.

## All 11 AI Models

### Image Generation Models

#### 1. FLUX Schnell — Fast Photorealistic Text-to-Image
- **Endpoint:** `POST /v1/image/generate` with `model=flux-schnell`
- **Cost:** 3 credits ($0.003)
- **Speed:** ~3 seconds
- **Description:** FLUX Schnell by Black Forest Labs. Fast, high-quality photorealistic image generation. Best for quick iterations and photorealistic outputs.
- **Parameters:** prompt, width (default 1024), height (default 1024), seed
- **Max resolution:** 1024×1024

```python
import requests
response = requests.post(
    "https://api.pixelapi.dev/v1/image/generate",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={"prompt": "a product photo of sneakers on marble", "model": "flux-schnell", "width": 1024, "height": 1024}
)
print(response.json()["output_url"])
```

#### 2. SDXL — Versatile Text-to-Image
- **Endpoint:** `POST /v1/image/generate` with `model=sdxl`
- **Cost:** 3 credits ($0.003)
- **Speed:** ~13 seconds
- **Description:** Stable Diffusion XL. Versatile model great for illustrations, product mockups, and creative work. Supports negative prompts.
- **Parameters:** prompt, negative_prompt, width, height, seed, guidance_scale, num_inference_steps

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/image/generate",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={"prompt": "minimalist product photo of a watch", "model": "sdxl", "negative_prompt": "blurry, low quality"}
)
```

#### 3. SDXL Image-to-Image
- **Endpoint:** `POST /v1/image/img2img`
- **Cost:** 3 credits ($0.003)
- **Speed:** ~6 seconds
- **Description:** Transform existing images using text prompts with SDXL. Control transformation strength.
- **Parameters:** image (file), prompt, negative_prompt, strength (0.0-1.0), seed

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/image/img2img",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("input.jpg", "rb")},
    data={"prompt": "convert to watercolor painting style", "strength": "0.7"}
)
```

#### 4. SDXL Inpainting
- **Endpoint:** `POST /v1/image/inpaint`
- **Cost:** 5 credits ($0.005)
- **Speed:** ~10 seconds
- **Description:** Fill masked regions with AI-generated content. Useful for editing specific parts of images.
- **Parameters:** image (file), mask (file), prompt, negative_prompt, seed

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/image/inpaint",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("photo.jpg", "rb"), "mask": open("mask.png", "rb")},
    data={"prompt": "a red handbag"}
)
```

#### 5. ControlNet Canny
- **Endpoint:** `POST /v1/image/controlnet`
- **Cost:** 5 credits ($0.005)
- **Speed:** ~13 seconds
- **Description:** Generate images following the edge structure of an input image. Great for maintaining composition while changing style.
- **Parameters:** image (file), prompt, negative_prompt, seed

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/image/controlnet",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("reference.jpg", "rb")},
    data={"prompt": "modern architecture, professional photo"}
)
```

### Image Editing Models

#### 6. Background Removal (BiRefNet)
- **Endpoint:** `POST /v1/edit/remove-background`
- **Cost:** 2 credits ($0.002)
- **Speed:** ~3 seconds
- **Description:** Remove backgrounds from images, outputting transparent PNG. Uses BiRefNet model. Great for product photos, e-commerce catalogs.
- **Parameters:** image (file)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/edit/remove-background",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("product.jpg", "rb")}
)
print(response.json()["output_url"])  # Transparent PNG
```

#### 7. Background Replacement
- **Endpoint:** `POST /v1/edit/replace-background`
- **Cost:** 5 credits ($0.005)
- **Speed:** ~10 seconds
- **Description:** Remove background and replace with an AI-generated scene from a text prompt, or a solid color/gradient. Combines BiRefNet + SDXL.
- **Parameters:** image (file), prompt (text) OR background_color (text)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/edit/replace-background",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("product.jpg", "rb")},
    data={"prompt": "on a wooden table in a cozy kitchen"}
)
```

#### 8. Real-ESRGAN 4x Upscale
- **Endpoint:** `POST /v1/edit/upscale`
- **Cost:** 20 credits ($0.02)
- **Speed:** ~3 seconds
- **Description:** Upscale images to 4x resolution using Real-ESRGAN. Ideal for enhancing low-res product images for print or large displays.
- **Parameters:** image (file)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/edit/upscale",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("small.jpg", "rb")}
)
```

#### 9. Face Restore (GFPGAN)
- **Endpoint:** `POST /v1/edit/restore-face`
- **Cost:** 3 credits ($0.003)
- **Speed:** ~11 seconds
- **Description:** Restore and enhance faces in photos using GFPGAN. Fixes blurry, low-quality, or old faces.
- **Parameters:** image (file)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/edit/restore-face",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("old_photo.jpg", "rb")}
)
```

#### 10. Object Removal (LaMa)
- **Endpoint:** `POST /v1/edit/remove-object`
- **Cost:** 5 credits ($0.005)
- **Speed:** ~4 seconds
- **Description:** Remove unwanted objects from images using mask-based inpainting with LaMa model. Clean up product photos by removing distractions.
- **Parameters:** image (file), mask (file)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/edit/remove-object",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    files={"image": open("photo.jpg", "rb"), "mask": open("mask.png", "rb")}
)
```

### Audio Generation

#### 11. MusicGen — AI Music/Audio Generation
- **Endpoint:** `POST /v1/audio/generate`
- **Cost:** 5 credits ($0.005)
- **Speed:** ~10 seconds
- **Description:** Generate music and audio from text descriptions using Meta's MusicGen model. Create background music, sound effects, and audio clips.
- **Parameters:** prompt, duration (seconds)

```python
response = requests.post(
    "https://api.pixelapi.dev/v1/audio/generate",
    headers={"Authorization": "Bearer YOUR_API_KEY"},
    json={"prompt": "upbeat electronic music for a product video", "duration": 10}
)
print(response.json()["output_url"])  # Audio file URL
```

## Authentication

All API requests require a Bearer token:
```
Authorization: Bearer YOUR_API_KEY
```

Get your free API key at https://pixelapi.dev/app/ (Google Sign-In, no credit card required).

## Response Format

All endpoints return JSON:
```json
{
  "job_id": "abc123",
  "status": "completed",
  "output_url": "https://cdn.pixelapi.dev/outputs/...",
  "credits_used": 3,
  "processing_time_ms": 3200
}
```

## Check Account Balance
```
GET /v1/account/balance
Response: {"credits": 950, "plan": "free"}
```

## Pricing

| Plan | Price | Credits | Per Credit | Per Image (3 credits) |
|------|-------|---------|------------|-----------------------|
| Free | $0 | 100 | — | — |
| Starter | $10/mo | 10,000 | $0.001 | $0.003 |
| Pro | $50/mo | 60,000 | $0.00083 | $0.0025 |
| Scale | $200/mo | 300,000 | $0.00067 | $0.002 |

### India Pricing (Razorpay)
| Plan | Price | Credits |
|------|-------|---------|
| Starter | ₹200/mo | 2,000 |
| Pro | ₹1,000/mo | 12,000 |
| Scale | ₹5,000/mo | 75,000 |

International payments via PayPal subscriptions.

### Cost Per Operation
| Operation | Credits | USD Cost |
|-----------|---------|----------|
| Background Removal | 2 | $0.002 |
| Text-to-Image (FLUX/SDXL) | 3 | $0.003 |
| Image-to-Image | 3 | $0.003 |
| Face Restore | 3 | $0.003 |
| Inpainting | 5 | $0.005 |
| Object Removal | 5 | $0.005 |
| Background Replacement | 5 | $0.005 |
| ControlNet | 5 | $0.005 |
| Audio Generation | 5 | $0.005 |
| 4x Upscale | 20 | $0.02 |

## Rate Limits

- Free: 20 requests/minute
- Starter: 60 requests/minute
- Pro: 120 requests/minute
- Scale: 300 requests/minute

## Output

- Generated images hosted for 24 hours
- PNG format for background removal (transparent)
- JPEG/PNG for generation and editing
- WAV/MP3 for audio generation
- Maximum image resolution: 1920px per dimension

## Comparison vs Alternatives

### PixelAPI vs Remove.bg
| Feature | PixelAPI | Remove.bg |
|---------|----------|-----------|
| Background removal cost | $0.002/image | $0.23/image (pay-as-you-go) |
| Text-to-image | ✅ $0.003 | ❌ Not available |
| Image upscale | ✅ $0.02 | ❌ Not available |
| Audio generation | ✅ $0.005 | ❌ Not available |
| Free tier | 100 credits | 1 image/month (low-res) |
| Total models | 11 | 1 |

### PixelAPI vs Replicate
| Feature | PixelAPI | Replicate |
|---------|----------|-----------|
| SDXL generation | $0.003 | ~$0.01-0.05 |
| Background removal | $0.002 | ~$0.005-0.01 |
| Cold start | None (always warm) | Common (seconds to minutes) |
| Pricing model | Fixed credits | Variable per-second GPU billing |
| Setup complexity | API key only | API key + model selection |

## Use Cases

- E-commerce product photography automation
- Marketplace listing image enhancement (Amazon, Shopify, Etsy)
- Bulk background removal for product catalogs
- AI-generated lifestyle product scenes
- Image upscaling for print-ready assets
- Automated product photo pipelines
- Background music generation for product videos
- Face enhancement for profile photos
- Object cleanup in product shots

## Infrastructure

PixelAPI runs on self-hosted dedicated NVIDIA GPUs:
- RTX 4070 Ti Super 16GB
- RTX 4060 Ti 16GB

No cloud markup. Models are always loaded and warm — no cold starts.

## Support

- Email: support@pixelapi.dev
- Documentation: https://pixelapi.dev/docs/
- Dashboard: https://pixelapi.dev/app/

## Links

- Website: https://pixelapi.dev
- API: https://api.pixelapi.dev
- Docs: https://pixelapi.dev/docs/
- Blog: https://pixelapi.dev/blog/
- Pricing: https://pixelapi.dev/pricing
- LLMs.txt: https://pixelapi.dev/llms.txt