ImageGen MCP
Free, keyless image generation MCP server that lets AI agents create images from text prompts via Pollinations/Flux. Supports multiple aspect ratios and returns local image file paths.
README
ImageGen MCP
Free, keyless image generation for any MCP-enabled AI agent. Give your assistant the ability to create images from text prompts — with no API key, no billing, and no account setup.
What it gives you
- One MCP tool:
generate_image— text prompt in, local image file out. Alsohealthfor readiness checks. - Zero configuration — no API keys, no accounts, no cost. The provider is Pollinations (Flux), a free community text-to-image service.
- Five aspect ratios —
1:1,3:4,4:3,9:16,16:9(default1:1). - Local-first output — images are saved to disk and returned as file paths; nothing is uploaded to a third party beyond the image request itself.
- Defensive by design — prompt length limits, output-dir writability checks, and structured
{ok, ...}/{ok: false, error}responses for clean agent error handling.
See it in action
Real outputs generated with this server (Pollinations / Flux, no API key, no post-processing):




Every image above was produced by a single generate_image call with just a text prompt and an aspect ratio.
How it works
AI agent → MCP tool generate_image(prompt, aspect_ratio) → Pollinations API → local image file → path returned to agent
The server speaks the Model Context Protocol (stdio transport) and uses FastMCP, so it plugs into any MCP host — including Hermes, Claude Desktop, or custom agent frameworks.
Installation
pip install -r requirements.txt # installs mcp
python imagegen_mcp_server.py # runs the MCP server over stdio
Or register it as a stdio MCP server in your host. Example (Hermes config):
mcp_servers:
imagegen:
command: /path/to/python
args:
- /path/to/imagegen_mcp_server.py
enabled: true
trust: full
Usage
Call the generate_image tool from your agent:
generate_image(prompt="a red sports car in a rainy city at night", aspect_ratio="16:9")
Returns on success:
{
"ok": true,
"path": "/home/user/gemini-image-mcp/output/img_20260816_222542_42026.jpeg",
"mime_type": "image/jpeg",
"size_bytes": 53907,
"aspect_ratio": "16:9",
"provider": "pollinations",
"message": "image saved to /home/user/gemini-image-mcp/output/img_20260816_222542_42026.jpeg"
}
Options
| Argument | Default | Notes |
|---|---|---|
prompt |
— (required) | Text description, max 2000 chars |
aspect_ratio |
1:1 |
One of 1:1, 3:4, 4:3, 9:16, 16:9 |
output_dir |
env or ~/gemini-image-mcp/output |
Where the image is saved; override per-call or via IMAGE_MCP_OUTPUT_DIR |
Notes & limitations
- Pollinations is a free community service — no SLA. Expect occasional slowness or brief downtime; fine for casual/prototyping use, not for a production dependency.
- The endpoint blocks default HTTP client user-agents (403), so the server sends a real browser
User-Agentheader. - Each image is roughly 20–65 KB, so storage stays negligible.
License
MIT — see LICENSE.
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。