Image MCP Server

Description

Image MCP Server is a Model Context Protocol server that lets any MCP-capable AI agent generate and edit images through a single connection. After pointing a client at the server once, an agent gains eight image tools it can call on its own, including generate_image, generate_transparent_image, edit_image, remove_background, upscale_image, text_to_svg, compress_image, and convert_format, with no per-provider SDKs or glue code.

It works with any client that speaks MCP, such as Claude Desktop and Claude.ai, Cursor, VS Code, Copilot, Windsurf, Zed, Codex, and more, and also offers an agent skill and a plain HTTP API with a Bearer token for backends and CI pipelines that do not use MCP. One key covers every model behind the server, producing a single bill instead of an account per provider, with bring-your-own-key supported on the Enterprise plan.

Each tool call costs a fixed number of credits, from 1 for a format conversion to 5 for a vector SVG and 8 to 30 for a generation depending on model tier. Plans refill credits monthly and credit packs never expire. A free tier provides 30 credits per month with no credit card. Outputs are returned as a public URL by default to keep agent context small, or as raw bytes on request, and each call is logged with prompt, model, latency, status, and credits deducted.

Image MCP Server's Core Features

• One MCP server exposing eight image tools to any agent

• Generate, edit, inpaint, and generate transparent images

• Background removal, upscaling, vectorization (text_to_svg)

• Image compression and format conversion

• Works with Claude, Cursor, VS Code, Copilot, and any MCP client

• Plain HTTP API with Bearer token for non-MCP stacks

• One key across 40+ models with a single bill

• Per-call request logging and automatic refunds handling

How to use Image MCP Server?

1. 1

   Connect the server: Point your MCP client at the server with one config line.

2. 2

   Use the tools: Let your agent call the eight image tools on its own.

3. 3

   Or call the API: Use the HTTP equivalent with a Bearer token for backend jobs.

4. 4

   Retrieve outputs: Get a public URL by default, or request raw bytes.

Image MCP Server's Use Cases

• Agent-built websites
• Unified image API
• Backend pipelines
• Spend auditing

FAQ from Image MCP Server

• What is an MCP server?

  Model Context Protocol is the open standard AI clients use to discover and call external tools. Point a client at this server once and your agent gains eight image tools it can call on its own, with no plugin code or per-provider SDK.

• Which agents and IDEs can connect?

  Anything that speaks MCP, including Claude Desktop and Claude.ai, Cursor, VS Code, Copilot, Windsurf, Zed, Codex, Antigravity, and more. There is also an agent skill and a plain HTTP API.

• Do I need my own provider keys?

  No. One key covers every model behind the server, so you get a single bill instead of an account per provider. Bring-your-own-key is supported on the Enterprise plan.

• How do credits work?

  Each tool call costs a fixed number of credits, from 1 for a format conversion to 5 for a vector SVG and 8 to 30 for a generation depending on the model tier. Credits refill monthly and credit packs never expire, with a free tier of 30 credits a month.

• Can I call it as a normal REST API?

  Yes. Every MCP tool has an HTTP equivalent that takes the same arguments and a Bearer token, so you can use it from a backend job or CI pipeline without an agent.

• Where do generated images live?

  Outputs are uploaded to object storage and returned as a public URL by default to keep agent context small, or you can pass responseFormat b64_json to receive raw bytes.