NeKo Biological Network Builder — Claude Code MCP server for pathway networks
Builds and analyses signalling networks from interaction databases. Useful when a model should come from known interactions instead of a blank sheet.
claude mcp add neko -- uvx mcp-biomodelling-serversWhat NeKo Biological Network Builder does
NeKo is one of four servers shipped in the mcp-biomodelling-servers package. It pulls interactions out of databases, assembles them into a signalling network, and then helps you inspect the result. The other three are MaBoSS, BioMASS and PhysiCell, each with its own job.
The package installs as a whole: python -m pip install mcp-biomodelling-servers. Modelling dependencies arrive with it, among them nekomata from 1.10.2 up to but not including 2.0, plus PyGraphviz. NeKo history diagrams render through PyGraphviz, whose wheels carry Graphviz, so a separate dot binary is not needed. Python 3.10–3.15 is required. There is also an isolated route: uvx --from mcp-biomodelling-servers mcp-neko-server.
The server speaks MCP over stdio, so in a terminal it sits waiting for a client on standard input. A package resolver set to reject every pre-release will fail the install, because the cmaboss engine exists only as 1.0.0b* builds.
What it helps with
- assemble a signalling network from interaction database records
- break a finished network down into nodes and edges
- wire biological modelling into an agent inside an editor
- run NeKo from an isolated environment without hand-editing configs
- draw a history diagram of a network
Example prompts for the agent
- Build a signalling network from the interactions in the database and list which nodes made it in
- Break this network down: which edges lead into the hub node
- Start NeKo with
uvx --from mcp-biomodelling-servers mcp-neko-serverand draw the network history diagram
How to install NeKo Biological Network Builder
Platform CLI
claude mcp add neko -- uvx mcp-biomodelling-serversuvx runs the command named after the package; if the README names another one, use uvx --from mcp-biomodelling-servers <command>
Platform CLI
codex mcp add neko -- uvx mcp-biomodelling-serversuvx runs the command named after the package; if the README names another one, use uvx --from mcp-biomodelling-servers <command>
Config file
{
"mcpServers": {
"neko": {
"command": "uvx",
"args": [
"mcp-biomodelling-servers"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from mcp-biomodelling-servers <command>
Config file
{
"servers": {
"neko": {
"type": "stdio",
"command": "uvx",
"args": [
"mcp-biomodelling-servers"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from mcp-biomodelling-servers <command>
Platform CLI
gemini mcp add neko uvx mcp-biomodelling-serversuvx runs the command named after the package; if the README names another one, use uvx --from mcp-biomodelling-servers <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about NeKo Biological Network Builder
How do I install NeKo Biological Network Builder in Claude Code?
Add the marketplace first: claude plugin marketplace add marcorusc/mcp-biomodelling-servers. Then install the plugin: claude plugin install neko@marcorusc. Add --scope project to the first command to declare it in the repository's .claude/settings.json.
What does NeKo Biological Network Builder do?
It builds signalling networks from database interactions and helps you analyse them. It runs as an MCP server over stdio.
What does NeKo need to run?
Python 3.10–3.15; the MCP Python SDK 2.x and modelling dependencies come with the package. No standalone Graphviz executable is required, since PyGraphviz bundles Graphviz in its wheels.
