DECIMER MCP Server
Turn chemical structure images into SMILES strings through DECIMER from your agent for cheminformatics notebooks, lab tooling, or research assistants.
Overview
DECIMER MCP Server is an MCP server for the Build phase that performs DECIMER image-to-SMILES chemical structure recognition for AI agents.
What is this MCP server?
- DECIMER image-to-SMILES chemical structure recognition exposed over MCP
- Python package decimer-mcp-server 0.1.4 via uvx on PyPI
- Stdio transport for chemistry-aware agent workflows
- Useful for digitizing diagrams from papers, slides, or lab scans
- Published version 0.1.4 on PyPI
- Stdio transport with uvx runtimeHint per server schema
What problem does it solve?
Your agent cannot interpret chemical structure drawings from images without a specialized recognition model wired into the toolchain.
Who is it for?
Builders automating chemistry research workflows, structure digitization, or edu tooling with MCP and Python uvx.
Skip if: General OCR needs, non-chemistry document parsing, or teams without Python/uvx tolerance on the agent machine.
What do I get? / Deliverables
With decimer-mcp-server on stdio, the agent returns SMILES strings from structure images for downstream cheminformatics scripts.
- SMILES strings extracted from chemical structure images via MCP
- Agent-ready cheminformatics recognition step in larger pipelines
Recommended MCP Servers
Journey fit
Specialized ML-backed recognition is integrated during build when you wire scientific pipelines into MCP workflows. integrations fits because the server wraps the DECIMER image-to-structure model as callable MCP tools, not general frontend or docs work.
How it compares
Domain-specific chemistry vision MCP, not a general vision LLM or PDF text extractor.
Common Questions / FAQ
Who is DECIMER MCP Server for?
Developers and researchers building MCP agents that need reliable image-to-SMILES conversion for chemical structures.
When should I use DECIMER MCP Server?
Use it during build when integrating lab, education, or cheminformatics pipelines that ingest structure images.
How do I add DECIMER MCP Server to my agent?
Register stdio MCP with runtimeHint uvx and package decimer-mcp-server 0.1.4 from PyPI, then invoke recognition tools on image inputs.