Metadata-Version: 2.4
Name: siliconflow-mcp
Version: 0.1.1
Summary: High-quality SiliconFlow Image Generation MCP Server
Author: stevefordev
Keywords: ai,flux,image-generation,mcp,qwen,siliconflow
Classifier: License :: OSI Approved :: MIT License
Classifier: Operating System :: OS Independent
Classifier: Programming Language :: Python :: 3
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.10
Requires-Dist: httpx>=0.28.1
Requires-Dist: mcp>=1.27.0
Requires-Dist: python-dotenv>=1.2.2
Description-Content-Type: text/markdown

# SiliconFlow MCP Server

[[English](README.md) | [한국어](README.ko.md)]

An MCP (Model Context Protocol) server for SiliconFlow's image generation service. This allows AI models (like Claude) to generate high-quality images directly using various models available on SiliconFlow.

## Features

- **`generate_image` tool**: Generate images from text prompts.
  - Supports multiple models (FLUX.1-schnell, FLUX.1-dev, FLUX.2-pro, etc.)
  - **`aspect_ratio` support**: Choose from 1:1, 16:9, 9:16, etc.
  - Supports `negative_prompt` for supported models.
  - Customizable seeds for reproducible generations.
- **`generate_video` tool**: Generate videos via text prompts (auto-polls until completion).
  - Supports Wan-AI models and customizable aspect ratios.
- **`submit_video_generation` & `get_video_status`**: Low-level tools for manual async video management.
- **`list_models` tool**: Dynamically fetch available image and video models.
- **`get_user_info` tool**: Check your SiliconFlow account details, including balance (Total, Paid, Free) and profile info.
- **Local Saving**: Automatically save `.png`, `.jpg`, or `.mp4` files to your specified directory.

## Setup

### 1. Prerequisites
- [uv](https://github.com/astral-sh/uv) installed (recommended) or Python 3.10+.
- A SiliconFlow API Key. Get one at [SiliconFlow Dashboard](https://cloud.siliconflow.com/account/ak).

### 2. Configuration
The server requires an API key to function. You can provide it via environment variables or a `.env` file.

```bash
SILICONFLOW_API_KEY=your_api_key_here
# Optional: Path to save generated images/videos locally
SILICONFLOW_IMAGE_DIR=C:/path/to/save/assets
```

## Usage

### Using with uvx (Recommended)
You don't need to install anything locally. Just run it directly using `uvx`:

```bash
uvx siliconflow-mcp
```

### Installation via PyPI
You can also install it as a global tool:

```bash
uv tool install siliconflow-mcp
# or using pip
pip install siliconflow-mcp
```

### Configuration for MCP Clients

#### Claude Desktop
Add the following to your Claude Desktop configuration file (`%APPDATA%\Claude\claude_desktop_config.json` on Windows or `~/Library/Application Support/Claude/claude_desktop_config.json` on macOS):

```json
{
  "mcpServers": {
    "siliconflow": {
      "command": "uvx",
      "args": ["siliconflow-mcp"],
      "env": {
        "SILICONFLOW_API_KEY": "your_api_key_here",
        "SILICONFLOW_IMAGE_DIR": "C:/path/to/save/assets"
      }
    }
  }
}
```

#### Claude Code
Run the following command:
```bash
claude mcp add siliconflow -- uvx siliconflow-mcp
```

#### Gemini CLI
Add the configuration to your `settings.json` (usually located in `.gemini/settings.json`):
```json
{
  "mcpServers": {
    "siliconflow": {
      "command": "uvx",
      "args": ["siliconflow-mcp"],
      "env": {
        "SILICONFLOW_API_KEY": "your_api_key_here",
        "SILICONFLOW_IMAGE_DIR": "C:/path/to/save/assets"
      }
    }
  }
}
```

## Installation for Developers
If you want to contribute or run from source:

```bash
# Install dependencies
uv sync

# Run the server locally
uv run siliconflow_mcp
```

## Supported Models
- `black-forest-labs/FLUX.1-schnell` (Fast and efficient)
- `black-forest-labs/FLUX.1-dev` (Higher quality)
- `black-forest-labs/FLUX.2-pro` (Professional grade)
- ...and other image models hosted on SiliconFlow.

## License
MIT
