Skip to content
Tresca

PyPSA MCP servers and AI agent skills

Two community MCP servers wrap PyPSA: PowerMCP's PyPSA server, which loads NetCDF or CSV networks and runs power flow, linear optimal power flow, capacity expansion and N-1 checks, and Open Energy Transition's PyPSA MCP, which builds small models from chat and optimises them. A PowerSkills skill covers planning studies. Both run in Python with an LP solver such as HiGHS and need no licence; the PyPSA developers publish no AI tooling.

  • No official Open source (TU Berlin) server yet.
  • 2 MCP servers and 1 agent skill listed.
Open source (TU Berlin)2 servers · 1 skillChecked 4 Nov 2026

Integrations for PyPSA

Community and commercial servers

MCP serverCommunity

PowerMCPby Power-Agent

One Python package with separate MCP servers for pandapower, PyPSA, OpenDSS, PowerWorld, PSS/E, PowerFactory, PSCAD and more.

LocalWindows, macOS, LinuxNeeds a PyPSA licenceCan modify models234· updated 7 days ago
MCP serverCommunity

PyPSA MCPby Open Energy Transition

Builds PyPSA energy models from chat: buses, generators, lines and storage, then power flow and optimisation runs.

LocalCan modify models74· updated 8 months ago

Agent skills

SkillCommunity

PowerSkillsby Power-Agent

Skills that tell an agent which PowerMCP tool to use first, plus ten mitigation playbooks for power-system study findings.

80· updated 13 days ago

Listings are sorted by origin, then by recent activity. GitHub data is refreshed nightly. Nothing here is ranked.

How AI connects to PyPSA

An MCP server or skill can only do what PyPSA lets programs do. PyPSA exposes 2 automation interfaces:

  • PyPSA Python API (Python): PyPSA is an MIT-licensed framework for optimising and simulating power and energy systems, with linearised AC and DC power flow, unit commitment, inter-temporal storage and coupling to other energy sectors. Optimisation calls an external LP solver; PowerMCP's server needs at least one, such as HiGHS, CBC, GLPK or Gurobi, and pins PyPSA between 0.35.2 and 2. Documentation
  • NetCDF and CSV network folders (Input files): Networks are stored as NetCDF (.nc) files or folders of CSV tables. PowerMCP imports and exports both, and its PowerIO layer converts PSS/E or MATPOWER cases into a NetCDF network, so a transmission case built elsewhere can be optimised in PyPSA.

What Open source (TU Berlin) offers

PyPSA's maintenance is led by TU Berlin's Department of Digital Transformation in Energy Systems, with many contributors. Its documentation describes the modelling framework and doesn't mention AI assistants or MCP. Open Energy Transition, which publishes the PyPSA MCP server, is a separate organisation, not the core PyPSA team.

Checked 4 Nov 2026. Sources: PyPSA documentation

What you can do with PyPSA and AI

  • Create a network with snapshots and add buses, generators, loads, lines and storage from a plain description, such as a two-bus 220 kV model.
  • Run AC or DC power flow, then linear optimal power flow with a chosen solver and formulation.
  • Run capacity expansion to size generation and storage, and summarise the optimised capacities.
  • Import a PSS/E or MATPOWER case into NetCDF, or load a CSV folder, and run an N-1 check.

Taken from the listings' own documentation. Check each listing for exactly what it supports.

Installing

The listings here document setup for Claude Code, Claude Desktop, Codex CLI and Cursor. Install steps, copied from each project's own documentation, are on each listing's page. Most PyPSA integrations run on your own computer next to PyPSA, so you need PyPSA installed and licensed on the same machine unless the listing says otherwise.

Related tools

pandapowerOpen source (University of Kassel and Fraunhofer IEE)1 server · 1 skillPowerWorld SimulatorPowerWorld1 server · 1 skillPSS/ESiemens1 server · 1 skillOpenDSSEPRI (open source)3 servers · 1 skill

Questions

Is there an official PyPSA MCP server?

No. The PyPSA MCP server comes from Open Energy Transition, and PowerMCP's server from the Power-Agent group; the PyPSA documentation doesn't list either.

Which PyPSA MCP server should I use?

PowerMCP's server works on existing networks (NetCDF, CSV or converted cases) and adds capacity expansion and N-1. Open Energy Transition's server is aimed at building small models from scratch in chat and doesn't document loading existing networks.

Do I need a commercial solver?

No. Open-source solvers such as HiGHS, CBC and GLPK work; PowerMCP's README recommends HiGHS and lists Gurobi as an option.

Can PyPSA results be compared with a PSS/E or pandapower run?

Yes, through PowerMCP. Its PowerIO layer converts a PSS/E .raw or MATPOWER case into a PyPSA NetCDF network and into pandapower, so the same case can be solved in each tool and the results compared.

AI-generated models, drawings and results must be checked by a qualified engineer before use.

Sources (checked 4 Nov 2026): PyPSA documentation · PowerMCP PyPSA server README · open-energy-transition/pypsa-mcp README · PyPSA product page

PyPSA is developed by its own project (Open source (TU Berlin)). Tresca is independent and not affiliated with or endorsed by the PyPSA project.