foxtrai-mcp

Foxtrai MCP Server Based on [CP](https://github.com/PHQ/fastmcp) framework, Foxtrai AI drawing platform MCP Server. ## Features - **upload_image** — Upload local image as reference image image - **create_d** — Submit AI drawing task - **get_task_status** — Query and result - **generate_image One-click generation (automatic polling) list_tasks** historical task list - **list — View image resource list delete_asset** — Delete image resource : `nano-banana`, `nano-banana-pro`, `ana-pro-ultra`, `nano-banana-2` ## Installation ```bash git clone git@github.com:liuruing/foxtrai-mcp.git cd foxtrai-mcp uv sync ``` ## Integrate with Claude Code Create `.mcp.json` in the project root directory: { "mcp { "foxtrai": "command": "uv", "args": ["run", "-- "/path/to/foxtcp", "python", "-mfoxtrai_mcp"], env": { "FOXTRAI_TOKEN": "your_token_here" } } } } ``` > Replace `/path/to/foxtrai-mcp` with the actual clone path, and `your_token_here` with your API Token. Restart Claude Code after configuration to use drawing-related tools in conversations. ## Integrate with Codex (OpenAI Codex CLI) Codex CLI also supports MCP Server. Add the following to `~/.codex/config.json`: ```json { "mcpServers": "fo": { "command":", "args": [" "--directory", "/path/to/foxtrai-mcp", "python", "-m", "foxtrai_mcp.server"], "env": { "FOXTRAI_TOKEN": "your_token_here" } } } } ``` ## Integrate with Cursor Create `.cursor/mcp.json` in the project root directory: ```json { "mcpServers": { "foxtrai": { "command": "uv", "args": ["run", "--directory", "/path/to/foxtrai-mcp", "python", "-m", "foxtrai_mcp.server"], "env": { "FOXTRAI_TOKEN": "your_token_here" } } } } ``` ## Run Directly ```bash # stdio mode (for MCP client connection) FOXTRAI_TOKEN=your_token_here uv run python -m foxtrai_mcp.server # SSE mode (HTTP service) FOXTRAI_TOKEN=your_token_here uv run python -c "from foxtrai_mcp.server import mcp; mcp.run(transport='sse', host='0.0.0.0', port=8000)" ``` ## Usage Examples After integration, you can directly say in Claude/Codex: - "Generate a cyberpunk-style city night scene with a 16:9 ratio using the nano-banana-pro model" - "Upload this image as a reference image, and then generate a new image with a similar style based on it" - "View the status of my recent drawing tasks" - "List all generated images" ## License MIT