@pipeworx/pjm

PJM Interconnection market data via Data Miner 2 — real-time LMPs, generation by fuel type and the seven-day load forecast for the largest US grid operator (13 states + DC, ~65 million people, Illinois to New Jersey).

Part of Pipeworx — an MCP gateway connecting AI agents to 1679+ live data sources.

Tools

• pjm_feeds(feed?, row_count?, start_row?) — list the Data Miner 2 feed catalog, or fetch ANY feed by exact name (da_hrl_lmps, inst_load, …). The escape hatch for the hundreds of PJM feeds without a dedicated tool here.

• pjm_lmp_realtime(pnode_id?, pnode_name?, datetime_beginning_ept?, row_count?) — five-minute real-time LMP with congestion and loss components.

• pjm_load_forecast(forecast_area?, row_count?) — seven-day hourly load forecast by transmission zone and RTO total.

• pjm_generation_fuel_mix(fuel_type?, datetime_beginning_ept?, row_count?) — MW and percentage of total by fuel type.

Related MCP server: aemo-mcp

Auth

BYO only today. Ocp-Apim-Subscription-Key header (Azure API Management). Platform fallback env var — not provisioned: PLATFORM_PJM_KEY.

A key is free but requires registering an account at https://apiportal.pjm.com/ and subscribing that account to the Data Miner 2 product. Those are two steps: a registered key that is not subscribed to the feed is refused identically to no key at all, which is why the pack's 401 message names the subscription explicitly rather than just saying "bad key".

Until the key lands, every tool refuses with a message containing requires an API key, and the pack is marked byoExpected: true in workers/gateway/src/pack-manifest.json so the vendor-key ratchet does not fail the deploy. Drop byoExpected the day PLATFORM_PJM_KEY is provisioned.

Data sources

• https://dataminer2.pjm.com/ — the browser UI over the same feeds.

• https://api.pjm.com/api/v1/ — the API. Documented envelope: { "totalRows": N, "items": [...], "links": [...] }.

Traps

• Response shapes here are UNVERIFIED. Holding no key, no PJM response body in this pack has been observed live. The pack normalizes the documented {totalRows, items, links} envelope and returns items as published, without renaming or reshaping individual fields. That is deliberate: a field mapping invented from documentation we could not exercise is a silent-wrong- answer waiting to happen, and returning upstream rows untouched is honest about what we actually know. Tighten this once a key exists and the real payload can be read.

• A 401 here tells you almost nothing. Verified 2026-09-17: api.pjm.com answers 401 with an empty body and no WWW-Authenticate header, both with no key and with a syntactically valid but unregistered one. So 401 cannot distinguish "no key" from "wrong key" from "key not subscribed to this feed". The pack says so in the error rather than guessing at a cause.

• format=json is not the default — Data Miner 2 defaults to another representation. This pack always sets it.

• Times are Eastern Prevailing Time (EPT — shifts with DST), and the range filter uses Data Miner's own "M/D/YYYY HH:MM to M/D/YYYY HH:MM" syntax, not ISO-8601.

• A 200 with 0 rows is treated as an error, not an empty success: it almost always means the date filter fell outside the feed's published range.

Quick Start

Add to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):

{
  "mcpServers": {
    "pjm": {
      "url": "https://gateway.pipeworx.io/pjm/mcp"
    }
  }
}

What this endpoint actually serves

tools/list at https://gateway.pipeworx.io/pjm/mcp returns the tools in the table above plus the shared Pipeworx meta-tools — ask_pipeworx, discover_tools, search_within, remember/recall and the rest of the gateway-wide set. So the tool count you see is larger than this table: a single-pack endpoint currently lists roughly 30 shared tools alongside the pack's own. The connection's initialize response states its exact scope, and is the authoritative answer for a given day.

This is deliberate, not multiplexing by accident. The meta-tools are what let a scoped connection answer a question this pack does not cover — via ask_pipeworx, which routes across the whole catalog — without you adding a second MCP server. There is currently no way to mount a pack endpoint without them; if the extra schemas cost you more context than the routing is worth, connect to the full gateway once rather than to several pack endpoints.

Or connect to the full Pipeworx gateway to get every pack's tools listed directly, instead of just this one's:

{
  "mcpServers": {
    "pipeworx": {
      "url": "https://gateway.pipeworx.io/mcp"
    }
  }
}

Both URLs reach the same gateway and the same 1679+ data sources. The only difference is which pack's tools are listed directly; ask_pipeworx reaches all of them from either one.

No MCP client? Call it over HTTP

curl -X POST https://gateway.pipeworx.io/v1/tools/pjm_feeds \
  -H 'Content-Type: application/json' \
  -d '{"row_count":50}'

No account needed for the first calls. Inspect any tool: GET https://gateway.pipeworx.io/v1/tools/pjm_feeds. Find one: POST https://gateway.pipeworx.io/v1/tools/search_packs with {"query":"..."}.

Standalone (no gateway account)

This package also runs as a local stdio MCP server — no Pipeworx account, no gateway round-trip:

{
  "mcpServers": {
    "pjm": {
      "command": "npx",
      "args": ["-y", "@pipeworx/mcp-pjm"]
    }
  }
}

Or run it directly to confirm it starts:

npx -y @pipeworx/mcp-pjm

It speaks MCP over stdin/stdout and answers initialize/tools/list/tools/call for only this pack's tools — none of the shared meta-tools the gateway connection above adds. Same source, same tools, no ask_pipeworx routing.

Using with ask_pipeworx

Instead of calling tools directly, you can ask questions in plain English — this works on the pack endpoint above as well as on the full gateway:

ask_pipeworx({ question: "your question about Pjm data" })

The gateway picks the right tool and fills the arguments automatically.

More

• Docs and guides

• pipeworx.io

License

MIT