README.md

---
title: PixVerse Video Generator API
emoji: 🎬
colorFrom: purple
colorTo: blue
sdk: docker
pinned: false
license: mit
---

# PixVerse Video Generation API

FastAPI application for generating videos from images using PixVerse AI, with storage on Digital Ocean Spaces.

## Features

- **Health Check Endpoint**: `/health` - Verify service status
- **Video Generation Endpoint**: `/generate-video` - Generate video from image
- Firebase authentication
- Hugging Face token authorization
- Digital Ocean Spaces storage

## Storage Structure

```
milestone/
└── valentine/
    ├── source/     # User uploaded images
    └── results/    # Generated videos
```

## Environment Variables

Create a `.env` file with:

```env
# PixVerse API Configuration
PIXVERSE_API_KEY=your-pixverse-api-key

# Digital Ocean Spaces Configuration
DO_SPACES_KEY=your-digital-ocean-spaces-key
DO_SPACES_SECRET=your-digital-ocean-spaces-secret
DO_SPACES_BUCKET=your-bucket-name
DO_SPACES_REGION=nyc3
DO_SPACES_ENDPOINT=https://nyc3.digitaloceanspaces.com

# Firebase Configuration
FIREBASE_PROJECT_ID=your-firebase-project-id

# Hugging Face Token (for authorization)
HF_TOKEN=your-hugging-face-token
```

## Installation

```bash
pip install -r requirements.txt
```

## Running Locally

```bash
python main.py
```

Or with uvicorn:

```bash
uvicorn main:app --host 0.0.0.0 --port 7860 --reload
```

## API Endpoints

### 1. Health Check

**GET** `/health`

Response:
```json
{
    "success": true,
    "message": "Service is running",
    "data": {
        "status": "healthy",
        "service": "PixVerse Video Generation API",
        "version": "1.0.0"
    }
}
```

### 2. Generate Video

**POST** `/generate-video`

**Headers:**
- `Authorization`: `Bearer <hf_token>`

**Form Data:**
- `image`: Image file (JPG, PNG, WebP)
- `user_id`: User ID string (optional)
- `prompt_text`: Prompt text for video generation (must contain keywords to match category)
- `firebase_id_token`: Firebase ID token
- `appname`: App name for database routing (optional)

**Category Matching (4 video types):**
The prompt_text is automatically matched to one of 4 categories:
- Contains "kiss" → Ai Kiss Video
- Contains "hug" → Ai Hug Video
- Contains "bf" or "boyfriend" → Ai BF
- Contains "gf" or "girlfriend" → Ai GF

**Success Response:**
```json
{
    "success": true,
    "message": "Video created successfully",
    "data": {
        "result_url": "https://your-bucket.nyc3.digitaloceanspaces.com/milestone/valentine/results/...",
        "source_image": "https://your-bucket.nyc3.digitaloceanspaces.com/milestone/valentine/source/...",
        "video_id": 12345
    }
}
```

**Error Response:**
```json
{
    "success": false,
    "message": "Error description",
    "data": null
}
```

## Hugging Face Deployment

The API is configured to run on port 7860 for Hugging Face Spaces deployment.

### Hugging Face Secrets

Set these secrets in your Hugging Face Space:
- `PIXVERSE_API_KEY`
- `DO_SPACES_KEY`
- `DO_SPACES_SECRET`
- `DO_SPACES_BUCKET`
- `DO_SPACES_REGION`
- `DO_SPACES_ENDPOINT`
- `FIREBASE_PROJECT_ID`
- `HF_TOKEN`

## cURL Example

```bash
curl -X POST "http://localhost:7860/generate-video" \
  -H "Authorization: Bearer your-hf-token" \
  -F "image=@/path/to/image.jpg" \
  -F "user_id=user123" \
  -F "prompt_text=A romantic kiss scene with beautiful effects" \
  -F "firebase_id_token=firebase-id-token" \
  -F "appname=collage-maker"
```