README.md

---
title: Antigravity API Proxy
emoji: "\U0001F680"
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
pinned: false
license: cc-by-nc-sa-4.0
---

# Antigravity API Proxy

A cloud-deployed API proxy service that converts Google AI Web sessions to standard API interfaces (OpenAI, Claude, Gemini formats).

## Features

- **Multi-Protocol Support**: OpenAI, Claude, and Gemini API formats
- **Token Management**: Automatic token refresh and rotation
- **Web UI**: Account and configuration management interface
- **Model Mapping**: Flexible model aliasing
- **API Key Authentication**: Secure access with multiple API keys

## Authentication

### Frontend Access (Web UI)
The web management interface is protected by HuggingFace Spaces password.
Set this in: **Settings → Access control → Password**

### API Proxy Authentication
Configure API keys in HuggingFace **Settings → Secrets**:

| Secret Name | Description |
|-------------|-------------|
| `API_KEYS` | Comma-separated list: `key1,key2,key3` |
| `API_KEY_1` | First API key |
| `API_KEY_2` | Second API key |
| `API_KEY` | Single API key (backward compatible) |

**Example:**
```
API_KEYS = sk-abc123,sk-def456,sk-ghi789
```

When calling the API, include your key in the request header:
```bash
# Option 1: Authorization header
curl -H "Authorization: Bearer sk-abc123" ...

# Option 2: X-API-Key header
curl -H "X-API-Key: sk-abc123" ...
```

## Endpoints

### OpenAI Compatible
- `POST /v1/chat/completions` - Chat completions
- `GET /v1/models` - List models

### Claude Compatible
- `POST /v1/messages` - Messages API
- `GET /v1/models/claude` - List Claude models

### Gemini Native
- `POST /v1beta/models/:model:generateContent` - Generate content
- `GET /v1beta/models` - List Gemini models

## Usage Examples

### OpenAI SDK (Python)
```python
from openai import OpenAI

client = OpenAI(
    base_url="https://your-space.hf.space/v1",
    api_key="sk-your-api-key"
)

response = client.chat.completions.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)
```

### Claude SDK (Python)
```python
import anthropic

client = anthropic.Anthropic(
    base_url="https://your-space.hf.space",
    api_key="sk-your-api-key"
)

message = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)
print(message.content)
```

### cURL
```bash
curl -X POST https://your-space.hf.space/v1/chat/completions \
  -H "Authorization: Bearer sk-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'
```

## Configuration

Access the web UI at the root URL to manage:
- **Accounts**: Add via Refresh Token
- **Model Mappings**: Map OpenAI/Claude models to Gemini
- **Proxy Settings**: Configure upstream proxy if needed

## Environment Variables

| Variable | Description | Required |
|----------|-------------|----------|
| `API_KEYS` | Comma-separated API keys | No* |
| `API_KEY_1`, `API_KEY_2`, ... | Individual API keys | No* |
| `PORT` | Server port (default: 7860) | No |

*If no API keys are configured, authentication is disabled.

## Security Notes

1. **Always set API keys** in production to prevent unauthorized access
2. **Enable Space password** to protect the management UI
3. Refresh tokens are stored in `/data` persistent storage
4. Consider the security implications of storing tokens in cloud environments