"Researching Effective Methods for Developing AI Agents" matching MCP servers:

Autonomous, carbon-aware building management system that pairs EnergyPlus digital twins with LLMs via the Model Context Protocol (MCP) for dynamic HVAC optimization and grid carbon reduction.

MCP server that provides live facility data access and operational tools, enabling the AI agent to query sensor readings, HVAC status, energy usage, alerts, and execute confirmed maintenance actions.

Enables MCP clients to query open data and a JSON API for the energy and CO2e cost of LLM tokens by model, GPU, and country.

MCP server providing tools for AI-agents to interact with Enapter EMS.

Enables AI assistants to search and analyze Turkish electricity and energy market regulation by scanning the Resmî Gazete and EPDK, EPİAŞ, ETKB and TEİAŞ announcements, extracting amended articles and effective dates, and ranking changes by market importance with reasoning. It also provides source health status and rule-based effective-date extraction from any text, without requiring an API key.

Exposes five deterministic tools over stdio so LLM agents can evaluate all seven kill methods, plan complete well kills and CWI stripping/snubbing operations, and create or step live simulator sessions without hallucinating safety-critical pressure math. Returns kill mud weight, MAASP, stage-by-stage procedures, rig gauges, kill sheets, and graded debriefs.

Enables safe, synthetic telecom event search, refinery blend validation, and legal provision lookup through MCP-style initialize, tools/list, and tools/call JSON-RPC methods.

MCP server for the Carbon Interface API (v1), enabling AI agents to estimate carbon emissions through natural language queries via Pipeworx gateway.

Enables AI assistants to fetch live environmental telemetry, verify local Fire TV and smart-meter hardware over TCP sockets, calibrate display and audio settings, orchestrate smart-home scenes, calculate energy savings, and answer natural-language questions about its capabilities.

Enables IDE agents to query PG&E's DRP Compliance ArcGIS FeatureServer for substations, feeders, ICA hosting capacity, 8760 load profiles, fire threat zones, and grid deficiencies through ten stdio tools.

Enables calculation of the most economical fuel (gasoline, ethanol, or natural gas) based on prices and vehicle consumption, following the 70% rule for ethanol and 60% for GNV.

Enables utility operators to forecast transformer oil temperature, assess synthetic outage risk, rank impacted assets, and generate maintenance plans through IBM Bob MCP tools.

Exposes EnergyPlus building simulation tools as MCP endpoints, enabling LLMs to run closed-loop energy optimization while maintaining occupant comfort.

Enables interaction with beestat thermostats via MCP tools, including generic API calls and convenience methods for reading thermostat, sensor, and runtime data.

Provides real-time electricity prices, cheapest hours, and contract comparison for 40+ countries, enabling AI agents to make energy-aware decisions.

Hourly electricity grid carbon intensity (gCO2eq/kWh) for 45 zones across Europe, the US and Great Britain, computed from ENTSO-E, EIA and NESO generation data. Four read-only tools; every value carries its interval timestamp and how stale it is. No API key, no account.

An unofficial read-only MCP server for Panasonic AiSEG2 home energy management controllers, enabling clients to read live power flow, per-circuit consumption, circuit names, daily totals, and historical energy/cost data from the device's local web interface.

Unofficial MCP server for the messdienst24.de utility portal, providing tools to retrieve heating and hot-water consumption data, download PDF reports, and compare usage trends.

Read-only MCP server for Home Assistant that computes statistical summaries, trends, and session detection from recorded sensor history, returning compact tables instead of raw data so small local models can answer questions about energy usage, climate, and appliance activity without token-heavy arithmetic.

An MCP server for US electricity demand and weather data that provides tools, resources, and prompts to query hourly demand, daily demand, peak demand, and demand vs temperature across regions, all running locally with a bundled SQLite database.