image-2-Agent
Allows generating images through an OpenAI-compatible image generation API.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@image-2-AgentGenerate a blue and white technology poster."
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Local Codex + MCP Image Agent Demo
这是一个本地运行的图片生成 Demo,完整调用链为:
React → FastAPI → Codex CLI → STDIO MCP generate_image → OpenAI-compatible 中转 API → 本地存储
项目只包含单用户本地原型需要的功能:提交提示词、异步查询状态、预览和下载图片。它不包含登录、计费、历史记录、图片编辑或多用户并发。
环境要求
Python 3.13+
Node.js 20+
已安装并登录的 Codex CLI
支持
POST /images/generations的 OpenAI-compatible 图片中转 API
Related MCP server: image-forge-mcp
安装
在项目根目录执行:
python -m venv .venv
.\.venv\Scripts\Activate.ps1
python -m pip install -e ".[dev]"
Copy-Item .env.example .env编辑 .env,至少填写:
IMAGE_API_BASE=https://relay.example.com/v1
IMAGE_API_KEY=your-local-key
IMAGE_MODEL=gpt-image-2真实 .env 已被 Git 忽略。不要把 Key 写进 .codex/config.toml、前端代码或日志。
安装前端依赖:
Set-Location frontend
npm install
Set-Location ..分层验证
先绕过 Codex 和 MCP 验证中转 API:
python scripts/test_relay_api.py成功后,图片会保存到 data/generated/test-output.<格式>。
项目级 .codex/config.toml 已配置 image_generator STDIO MCP。项目被 Codex 信任并激活虚拟环境后,可检查连接:
codex mcp list
codex进入 Codex TUI 后使用 /mcp,确认存在 image_generator 和 generate_image。也可以要求:
使用 image generator MCP 工具生成一张蓝白色科技风海报。启动
终端一,启动后端:
激活后,提示符前会出现 (.venv)
python -m pip install -e ".[dev]"
启动后端
python -m uvicorn backend.app.main:app --host 127.0.0.1 --port 8000 --reload
python -m uvicorn backend.app.main:app --host 127.0.0.1 --port 8000 --reload终端二,启动前端: Set-Location d:\project\image-2-Agent\frontend
Set-Location frontend
npm run dev浏览器打开 http://127.0.0.1:5173。
API
POST /api/generations:创建任务,返回job_idGET /api/generations/{job_id}:查询任务状态GET /api/images/{image_id}:预览图片GET /api/images/{image_id}/download:下载图片GET /api/health:健康检查
任务状态保存在内存中,服务重启后会丢失。图片和元数据分别保存在 data/generated 与 data/results。
测试与构建
pytest
Set-Location frontend
npm run buildProvider 和后端测试全部使用 mock,不会请求真实中转 API,也不会产生图片费用。真实端到端测试必须由本机已登录的 Codex 和有效中转 Key 完成。
常见错误
codex_not_found:后端账号找不到 Codex CLI。codex_not_authenticated:先在同一操作系统账号下登录 Codex。mcp_unavailable:检查虚拟环境、项目信任和.codex/config.toml。provider_auth_failed:检查 API Base、Key 和模型别名。provider_rate_limited:请求过于频繁,稍后重试。provider_quota_exhausted:中转 API 免费额度已用完,请充值或更换模型。provider_unavailable(no_available_channel):免费模型通道暂不可用,可改用gpt-image-2或稍后重试。provider_auth_failed:检查 API Key 是否来自当前平台,修改.env后需重启后端。provider_timeout:提高IMAGE_API_TIMEOUT_SECONDS或检查中转服务。tool_not_called:Codex 没有实际调用generate_image,结果会被拒绝。mcp_tool_cancelled/user cancelled MCP tool call:Codexexec在read-only沙箱下会取消 MCP 工具调用。将.env中CODEX_SANDBOX设为danger-full-access。provider_invalid_response(404 Invalid URL):IMAGE_API_BASE应填写到/v1,不要包含/images/generations。
This server cannot be deployed
Maintenance
Related MCP Connectors
Focused MCP server for OpenAI image/audio generation (v2.0.0). Wraps endpoints via HAPI CLI.
MCP server for Qwen Image 3 AI image generation
MCP server for Flux AI image generation
MCP server for Wan AI video generation
Related MCP Servers
- AlicenseBqualityDmaintenanceA local MCP server for generating and editing images using OpenAI-compatible APIs. It provides text-to-image generation and image editing capabilities with configurable endpoints and saves output directly to local files.23 npmMIT
- FlicenseNot gradedqualityDmaintenanceAn MCP server for AI-powered image processing (generate, edit, vary, analyze) supporting OpenAI, Gemini, Ideogram, and custom relay endpoints.-
- AlicenseNot gradedqualityCmaintenanceA minimal MCP server for generating images via OpenAI's GPT image model, supporting inline display or file output.MIT
- AlicenseNot gradedqualityBmaintenanceMCP server that enables Codex to run image generation and inspection as durable jobs, returning bounded text results and references instead of image bytes to keep context manageable.100Apache 2.0