mcp-simple-image

One tool, one endpoint. simple_image posts a prompt to a StepFun-compatible /images/generations API and returns the resulting image URLs — no SDK, no third-party packages.

一个工具、一个端点:把提示词发给 StepFun 兼容的图片生成 API,拿回图片 URL。

As a DeepSeek Harness plugin: the MCP server ships inside the bundle, so installing one plugin is the whole setup — no mcpServers file to hand-edit.

Install

DeepSeek Harness Desktop — open Plugins in the sidebar, choose Add plugin, and enter:

https://github.com/bauerelizabeth07139/mcp-simple-image

Then switch the new dsh-simple-image bundle on. The Desktop app boots the reserved desktop profile, so that is where it has to be enabled.

dsh CLI — install it into the profile you actually boot:

dsh plugin --profile web add bauerelizabeth07139/mcp-simple-image

No git on the machine? pnpm resolves a git shorthand with git ls-remote, which fails with 'git' is not recognized when git is missing. Use the tarball instead — that path is plain HTTPS:

dsh plugin --profile web add https://codeload.github.com/bauerelizabeth07139/mcp-simple-image/tar.gz/main

The same address works in the Desktop Add plugin dialog. Replace main with a commit SHA to pin an exact revision (/tar.gz/<sha>).

Uninstall with dsh plugin --profile web remove dsh-simple-image.

Related MCP server: opencode-image-mcp

Requirements

• Python ≥ 3.8 on PATH, or pointed at with python.

• No third-party packages — the server is pure standard library.

• A StepFun-compatible API key in STEP_API_KEY (or config.apiKey). Without one the server exits at startup with ERROR: STEP_API_KEY is not set.

• An endpoint that already includes /v1 — the server appends /images/generations itself and hardcodes response_format: url.

Tools

The server registers 1 tool(s). DSH namespaces them automatically, so the model calls them as mcp__simple_image__<tool>:

Configuration

Every field is optional and lives in the loader row. For example, in cordis.patch.yml:

- id: dsh-simple-image
  name: 'dsh-simple-image'
  config:
    apiKey: 'sk-...'
    baseUrl: 'https://api.stepfun.com/step_plan/v1'
    model: 'step-image-edit-2'

Why this plugin ships a bridge

The server frames its stdio transport the LSP way — Content-Length: <n> followed by a blank line — while the MCP client inside DeepSeek Harness reads newline-delimited JSON. Mounted directly, the handshake would time out with no error worth reading. bridge.mjs sits between the two: NDJSON in from the harness, Content-Length frames to the server, and the reverse on the way back.

Messages are never re-encoded, only reframed, and the parser tolerates a \r\r\n\r\r\n header — which is what this server emits on Windows, where Python's text mode turns each \n into \r\n.

Notes

• config.json wins over the environment. If config.json (or STEP_CONFIG) exists next to the server, its api_key / api_base / model values are read first. Do not leave a copied config.example.json in place: its placeholder key is truthy, so the server starts with a bogus credential.

• Results are URLs, not files. The endpoint is asked for response_format: url, so the tool returns links that expire; nothing is written to disk.

• The base URL must already contain /v1. The server appends /images/generations verbatim, so an OpenAI-style base URL without /v1 produces a 404.

How it is mounted

index.js resolves a Python interpreter (the configured python, then python3/python on PATH), hands the server its argv and working directory, and mounts it as a stdio MCP server through @deepseek-ai/dsh-mcp-client with failOnStartupError: true, so a server that cannot start is a visible error rather than a silently missing tool.

Credentials are forwarded explicitly. The harness scrubs credential-shaped variables (KEY, TOKEN, SECRET, PASSWORD) out of the environment a child process inherits, so config.apiKey — falling back to the variable the server documents — is written into the child's environment by the plugin itself. That means both of these work:

config:
  apiKey: '<your key>'

export STEP_API_KEY='<your key>'   # picked up at load time

Development

No build step and no runtime dependencies — @deepseek-ai/cordis and @deepseek-ai/dsh-mcp-client are peers supplied by the Harness.

npm test    # node >= 22: manifest checks + the stdio mount, both Harness-free

The mount test loads index.js with @deepseek-ai/dsh-mcp-client stubbed and asserts the exact stdio configuration the plugin produces, including the credential forwarding above.

Repository layout

Other hosts (unchanged)

The server is a plain stdio MCP server and still works anywhere else. The repository's original README is kept verbatim as README.opencode.md, and the launch stanza from it keeps working:

STEP_API_KEY=sk-... python server.py                       # any MCP host that
                                                           # frames with
                                                           # Content-Length
node bridge.mjs -- python server.py                        # …or line-delimited
                                                           # JSON, through the
                                                           # bridge

License

MIT — the repository declared MIT in package.json and its README but shipped no licence file; this plugin's release adds one.