Purpose
SEOSiri Bio-Robotics Core Engine is an open-source, stateless Model Context Protocol (MCP) framework that acts as a stateless agent, bridging the gap between biological datasets, physiological biosignals, and physical robot actuation systems. It ingests biological inputs dynamically fetched from standard Web APIs (such as UniProt) or fallback reference data, as well as physiological biosignals (such as surface electromyography, EMG), normalizes them through mathematical bounds, and translates them into Cartesian physical coordinates expressed explicitly in International System of Units (SI, meters), which are then mapped to standardized G-code strings for physical motion planning or robotic prosthesis actuation.
Key Capabilities
This server exposes six MCP tools.
The framework has the following technical characteristics:
- Communication protocol: Model Context Protocol JSON-RPC, based on stdio.
- Driving protocol: Cartesian G-code, following the NIST RS274/NGC interpreter specification.
- Coordinate space: International System of Units (meters), with millimeter-level conversion provided for board-level operations.
- Validation: Automated unit testing via pytest.
- Containerization: Includes a Dockerfile for isolated runtime environments.
- Directory synchronization: glama.json drives automatic synchronization of the Glama.ai listing.
- Language: Python.
- Compatible platforms: Linux, macOS, Windows.
- Dependencies: uv (optional, for zero-configuration remote execution) or pip; and an MCP-compatible client (Claude Desktop, Cursor IDE, or Glama's browser-based Inspector).
The codebase is modular and readable, with a clear separation of concerns (core_math.py, hardware_gateway.py, main_mcp_server.py), and can serve as a reference implementation for building your own physical hardware MCP server.
Some tools have explicit safety and validation capabilities:
- translate_emg_to_actuation is a translation and safety check layer suitable for research and prototyping, and is not a certified medical or prosthetic control system. Any clinical or assistive deployment requires independent regulatory validation and is beyond the scope of this project.
- Closed-loop telemetry: parses real-time coordinate position queries from the serial port to verify physical-toAchieved.
- Capacitive liquid level detection (LLD): stops movement immediately when the probe contacts the liquid surface, preventing pipette tip damage.
All six tools have been independently tested by Glama.ai using the browser-based MCP Inspector against a running server, and the outputs match the closed-form mathematical model of each tool exactly.
How to Use
Quick Start (CLI Orchestrator)
The project uses a local-first Python runtime to bypass IDE proxy latency and regional API restrictions, providing a stable, deterministic execution channel.
Install the package in editable mode:
Run the decoupled orchestrator (live UniProt fetching):
Run with the fallback database:
Validate the CI/CD pipeline:
Connecting to Claude Desktop or Cursor IDE
This server can be connected to a local AI client using one of two standard methods.
Method 1: Execute directly from GitHub (zero configuration, recommended)
If uv is installed, the server can be run directly from the public repository without cloning it locally. Open claude_desktop_config.json (Windows: %APPDATA%\Claude\claude_desktop_config.json | macOS: ~/Library/Application Support/Claude/claude_desktop_config.json) and add the configuration.
Method 2: Local execution (if already cloned)
If the repository has been cloned to a local drive (e.g., D:/my-century-biorobotics-core), configure the client to point to the local entry file.
Applicable Scenarios
This framework is suitable for research and prototyping scenarios that combine biological datasets with physical robot actuation, such as robot motion planning based on biological data or physiological signals, robotic prosthetic actuation, and automation tasks that require converting biological signals into standardized physical coordinates. Its stateless agent architecture and modular code also make it suitable as a reference implementation for building custom physical hardware MCP servers.
Limitations
translate_emg_to_actuation is only a translation and safety-check layer, suitable for research and prototyping, and is not a certified medical or prosthetic control system. Any clinical or assistive deployment would require independent regulatory validation, which is beyond the scope of this project. The project is distributed under the MIT license.
