decimer-mcp-server 0.1.4

MCP adapter server for DECIMER FastAPI image-to-SMILES endpoint

DecimerMCPServer

mcp-name: io.github.DocMinus/decimer-mcp-server

MCP server that exposes DECIMER image-to-SMILES functionality as tool calls.

This project is a thin adapter over the existing FastAPI service in DecimerServerAPI. It does not run DECIMER models directly. The adapter sends JSON requests by default, with automatic fallback to form payloads for compatibility.

Tools

• server_health: Checks whether the DECIMER FastAPI server is reachable.
• analyze_chemical_image: Sends a base64-encoded image to /image2smiles/ and returns structured output.

Requirements

• Python 3.10+
• Running DECIMER API server (default: http://localhost:8099)

find it at either of these two versions:

• GitHub: https://github.com/DocMinus/DecimerServerAPI
• Dockerhub: https://hub.docker.com/r/docminus/decimer_api

Install

cd /Users/a/dev/DecimerMCPServer
uv venv
uv sync

Configuration

Copy .env.example values into your environment:

• DECIMER_API_BASE_URL (default http://localhost:8099)
• DECIMER_API_TIMEOUT_SECONDS (default 60)
• DECIMER_MAX_IMAGE_BYTES (default 10000000)
• DECIMER_MCP_LOG_LEVEL (default INFO)

Run (stdio transport)

uv run decimer-mcp-server

or

uv run python -m decimer_mcp_server

Example MCP client config

{
  "mcpServers": {
    "decimer": {
      "command": "uv",
      "args": ["run", "python", "-m", "decimer_mcp_server"],
      "env": {
        "DECIMER_API_BASE_URL": "http://localhost:8099"
      }
    }
  }
}

Output shape

analyze_chemical_image returns:

{
  "ok": true,
  "smiles": "CCO",
  "reason": null,
  "api_status_code": 200,
  "api_message": null,
  "classifier_score": 0.0000012,
  "classifier_threshold": 0.3,
  "classifier_decision": "structure_like"
}

When no SMILES is returned by API classifier behavior:

{
  "ok": true,
  "smiles": null,
  "reason": "not_chemical_structure",
  "api_status_code": 200,
  "api_message": "No SMILES returned by API",
  "classifier_score": 0.99999,
  "classifier_threshold": 0.3,
  "classifier_decision": "not_structure_like"
}

Development tests

uv sync --extra dev
uv run pytest

Make targets:

make sync
make test

Smoke test helper

Run one health check + one inference call against your DECIMER API:

cd /Users/a/dev/DecimerMCPServer
DECIMER_API_BASE_URL=http://chitchat:8099 uv run decimer-mcp-smoke-test --image /Users/a/dev/DecimerServerAPI/example_usage/structure.png

If you keep settings in .env, load it with:

uv run --env-file .env decimer-mcp-smoke-test --image /Users/a/dev/DecimerServerAPI/example_usage/structure.png

or use make:

make smoke

Override the image path if needed:

make smoke SMOKE_IMAGE=/absolute/path/to/image.png

## MCP Registry publishing

Tags matching `v*` trigger `.github/workflows/publish-mcp.yml`.

Workflow steps:
- installs `mcp-publisher`
- validates `server.json`
- calls registry publish using secret `MCP_REGISTRY_TOKEN`
- publishes slug `io.github.DocMinus/decimer-mcp-server` (case sensitive; must match registry grant)

Before tagging:
1. Update `pyproject.toml` + `server.json` versions
2. Ensure `server.json` stays valid (`uv pip install jsonschema && python validate snippet from AGENTS.md`)
3. Add GitHub repo secret `MCP_REGISTRY_TOKEN` (GitHub PAT with `repo`, `workflow` scopes)

Release flow:
```bash
git tag v0.1.1
git push origin v0.1.1

Monitor Actions tab. If publish fails, rerun using workflow dispatch after fixing issues.

## Contribution
This project was built by DocMinus with AI-assisted coding support (OpenCode/Copilot-style tooling), then reviewed and tested by the author.

## AI usage policy

- AI assistance was used for scaffolding, implementation drafts, and documentation edits.
- Final technical decisions, validation runs, and acceptance were performed by the maintainer.
- Runtime behavior should be validated with local tests (`make test`) and smoke tests (`make smoke`) before release.