bfl-flux-mcp
Minimal Python MCP server for Black Forest Labs Flux image generation API, providing tools for text-to-image, editing, expansion, variations, and credit checking.
README
BFL Flux MCP Server
Minimal Python MCP server for Black Forest Labs Flux image generation API.
Features
-
Seven comprehensive tools:
generate_image- Text-to-image with model selectionedit_image- Image editing with natural languageexpand_image- Directional outpainting (add pixels to any side)create_variation- Create variations of an image using Redux (image_prompt)save_image- Download and save images before URLs expirecheck_credits- Verify API key and credit balancelist_finetunes- View your custom finetuned models
-
All Flux models supported:
- Generation:
flux-pro-1.1(default),flux-pro-1.1-ultra,flux-2-pro,flux-2-flex,flux-2-max,flux-pro,flux-dev - Editing:
kontext-pro,kontext-max,fill-pro - Expansion:
expand-pro
- Generation:
-
Full API parameter support:
- Aspect ratio or custom dimensions
- Prompt upsampling for enhanced output
- Output format (PNG/JPEG)
- Safety tolerance levels
- Model-specific: guidance, steps, raw mode
Installation
# With uv (recommended)
uv sync
# With pip
pip install -e .
Configuration
Get your API key at api.bfl.ml
export BFL_API_KEY="your-api-key-here"
Claude Code / MCP Config
Add to .mcp.json:
{
"mcpServers": {
"bfl-flux": {
"command": "uv",
"args": ["run", "--project", "/path/to/bfl-flux-mcp", "bfl-flux-mcp"],
"env": {
"BFL_API_KEY": "${BFL_API_KEY}"
}
}
}
}
Or with uvx (once published):
{
"mcpServers": {
"bfl-flux": {
"command": "uvx",
"args": ["bfl-flux-mcp"],
"env": {
"BFL_API_KEY": "${BFL_API_KEY}"
}
}
}
}
Usage
Check Credits
Verify your API key is working and check balance:
Check my BFL credits
Generate Image
Generate an image of a mountain landscape at sunset, photorealistic style
Parameters:
| Parameter | Description | Default |
|---|---|---|
prompt |
Text description (required) | - |
model |
Model to use | flux-pro-1.1 |
aspect_ratio |
1:1, 16:9, 9:16, 4:3, 3:4, 21:9, 9:21 |
1:1 |
width/height |
Custom dimensions (multiple of 16, max 2048) | - |
seed |
For reproducibility | - |
safety_tolerance |
0 (strict) to 6 (permissive) | 2 |
prompt_upsampling |
Enhance prompt for richer output | false |
output_format |
png or jpeg |
png |
guidance |
Prompt adherence 1.5-10 (flux-2-flex only) | - |
steps |
Generation steps 1-50 (flux-2-flex only) | - |
raw |
Raw mode (flux-pro-1.1-ultra only) | false |
Edit Image
Edit this image to add a rainbow in the sky
Parameters:
| Parameter | Description | Default |
|---|---|---|
prompt |
Edit instructions (required) | - |
image |
Base64-encoded image or URL (required) | - |
model |
kontext-pro, kontext-max, fill-pro |
kontext-pro |
aspect_ratio |
Output aspect ratio | preserve |
seed |
For reproducibility | - |
safety_tolerance |
0-2 | 2 |
Expand Image
Expand an image by adding pixels to any side (outpainting):
Expand the image by 256 pixels on the right side
Parameters:
| Parameter | Description | Default |
|---|---|---|
image |
Image to expand: path, URL, or base64 (required) | - |
prompt |
Optional guidance for the expansion | - |
top |
Pixels to add to top (0-2048) | 0 |
bottom |
Pixels to add to bottom (0-2048) | 0 |
left |
Pixels to add to left (0-2048) | 0 |
right |
Pixels to add to right (0-2048) | 0 |
seed |
For reproducibility | - |
safety_tolerance |
0-6 | 2 |
output_format |
png or jpeg |
png |
Save Image
Download and save a generated image before the URL expires (10 min):
Save that image to /path/to/logo.png
Parameters:
| Parameter | Description | Default |
|---|---|---|
url |
Image URL from generate/edit result (required) | - |
path |
Destination file path (required) | - |
List Finetunes
View your custom finetuned models:
Show my finetuned models
No parameters required.
Create Variation
Create variations of an existing image that maintain its essence while applying optional text guidance:
Create a variation of this image with a warmer color palette
Parameters:
| Parameter | Description | Default |
|---|---|---|
image |
Source image: path, URL, or base64 (required) | - |
prompt |
Optional text guidance for the variation | - |
model |
flux-pro-1.1, flux-pro-1.1-ultra, flux-dev |
flux-pro-1.1 |
aspect_ratio |
Output aspect ratio | 1:1 |
seed |
For reproducibility | - |
safety_tolerance |
0-6 | 2 |
output_format |
png or jpeg |
png |
Pricing
| Model | Credits | USD |
|---|---|---|
| flux-pro-1.1 | 4 | $0.04 |
| flux-pro-1.1-ultra | 6 | $0.06 |
| flux-2-pro | 5 | $0.05 |
| flux-2-flex | 1-5 | $0.01-0.05 |
| flux-2-max | 6 | $0.06 |
| kontext-pro | 4 | $0.04 |
| kontext-max | 8 | $0.08 |
| expand-pro | 4 | $0.04 |
1 credit = $0.01 USD
Note: Credits used are shown in tool responses when available.
Development
# Install with dev dependencies
uv sync --extra dev
# Run linter
uv run ruff check .
# Format code
uv run ruff format .
# Run tests
uv run pytest -v
# Run tests with coverage
uv run pytest --cov=bfl_flux_mcp --cov-report=term-missing
License
MIT - see LICENSE
Links
- BFL API Documentation
- BFL Playground (free testing)
- GitLab Repository
- GitHub Mirror
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。