EPW.tools is an MCP server with 9 tools an agent can call, among them analyze_epw, generate_epw, generate_forecast. It answers at https://epw.tools/api/mcp over streamable HTTP, and it listed its tools to an anonymous client, so no sign-in was needed for that. Our probes reached it 3 of 4 times since 29 September 2026; the last check, on 15 November 2026, succeeded.
EnergyPlus weather files (EPW, DDY, STAT) and design conditions for any site, OpenStudio-MCP ready.
Description as the publisher wrote it.
| Tool | What it does |
|---|---|
analyze_epw | Summary statistics of any public EPW by URL (OneBuilding, EnergyPlus, a shared file): location header, dry-bulb mean/min/max, annual global horizontal irradiation, data range warnings. Free, no key. 10 MB max. Inputs: url |
generate_epw | Start an actual-year (AMY) EPW + DDY + STAT for any point, 1990 to last year. 1 credit; +2 with urban_street_level (53 US cities: hourly street-level air temperature from citizen stations). station_id (from get_nearest_stations) builds the Inputs: lat, lon, year, api_key, location_name, data_source |
generate_forecast | Start a forecast EPW for the next 1-15 days (operational sims, MPC, demand response). 1 credit; more than 7 days needs an account that has bought credits. Inputs: lat, lon, api_key, days, location_name |
generate_future_climate | Start a CMIP6-morphed future EPW (Belcher 2005). scenario: ssp126, ssp245, ssp370, ssp585; horizon: any decade 2030-2100; model: leave empty for the default, the CMIP6 ensemble median (per month and variable, the median change of the ten CM Inputs: lat, lon, api_key, scenario, horizon, location_name |
generate_tmy | Start a typical meteorological year from a chosen period (at least 5 years; 10-20 is usual). 1 credit. tmy_method: "nrel" (Sandia method, default) or "onebuilding" (ISO 15927-4; paid). data_source: "hybrid" (station weather + reanalysis sun Inputs: lat, lon, start_year, end_year, api_key, location_name |
get_design_conditions | Heating and cooling design conditions for any site, free, in one of two modes that are never mixed. mode="ashrae_method": the ASHRAE Handbook Ch. 14 method replicated at the nearest ASHRAE station for the chosen edition (2025, 2021, 2017, 2 Inputs: lat, lon, mode, edition, current_climate, location_name |
get_job_result | Status of a job started by a generate_* tool; when complete, the download URLs. The links are valid for a short time (see downloads_note); call this tool again for fresh ones. Inputs: job_id, api_key |
get_nearest_stations | Nearest weather stations to a point (name, WMO id, distance, years of record). Free, no key. meteostat_hourly_end (also given as hourly_end) is the end date of the Meteostat station inventory, not the station's last report: newer hours are Inputs: lat, lon, n |
list_cmip6_models | The models available for future-climate morphing: `default` is what a request without a model gets (the CMIP6 ensemble median, id "ensemble-median"); `models` are global CMIP6 models (NASA NEX-GDDP, 25 km); `regional` are EURO-CORDEX pairs |
Every block below points at https://epw.tools/api/mcp. Pick your client and copy.
claude mcp add --transport http epw-tools https://epw.tools/api/mcp
{
"mcpServers": {
"epw-tools": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://epw.tools/api/mcp"]
}
}
}{
"mcpServers": {
"epw-tools": {
"url": "https://epw.tools/api/mcp"
}
}
}{
"servers": {
"epw-tools": {
"type": "http",
"url": "https://epw.tools/api/mcp"
}
}
}{
"context_servers": {
"epw-tools": {
"url": "https://epw.tools/api/mcp"
}
}
}{
"mcpServers": {
"epw-tools": {
"serverUrl": "https://epw.tools/api/mcp"
}
}
}{
"mcpServers": {
"epw-tools": {
"url": "https://epw.tools/api/mcp"
}
}
}name: EPW.tools
version: 0.0.1
schema: v1
mcpServers:
- name: EPW.tools
type: streamable-http
url: https://epw.tools/api/mcp{
"mcpServers": {
"epw-tools": {
"httpUrl": "https://epw.tools/api/mcp"
}
}
}{
"mcpServers": {
"epw-tools": {
"type": "http",
"url": "https://epw.tools/api/mcp"
}
}
}{
"mcpServers": {
"epw-tools": {
"url": "https://epw.tools/api/mcp"
}
}
}goose://extension?url=https%3A%2F%2Fepw.tools%2Fapi%2Fmcp&type=streamable_http&id=epw-tools&name=EPW.tools
It answered our last check on 15 November 2026 in 626 ms. Over 4 checks since 29 September 2026 it answered 3. Status here comes from Neuronto's own probes, not from the publisher.
9, as returned by the server's own tools/list on 18 November 2026: analyze_epw, generate_epw, generate_forecast, generate_future_climate, generate_tmy, get_design_conditions, get_job_result, get_nearest_stations, and 1 more.
It answered our last check on 15 November 2026 in 626 ms. Over 4 checks since 29 September 2026 it answered 3.
Put the server under mcpServers in ~/.cursor/mcp.json with the URL https://epw.tools/api/mcp. The exact JSON is above, with the same setup for Claude Code, Claude Desktop, VS Code, Windsurf, Zed, Gemini CLI and the rest.
It listed its tools to an anonymous client, so reading the tool list needs no key. Individual tools can still ask for credentials in their own inputs.
Publisher: io.github.carlobianchi89. Endpoint host: epw.tools. ARD identifier: urn:air:io.github.carlobianchi89:mcp:epwtools.
First indexed 29 September 2026. The tools are exactly what the server returned; nothing on this page
is inferred from its name. Ranking in Neuronto search is never for sale.