Prospector Labs · energy infrastructure API

Every interconnection queue in America, one call away.

50,597 interconnection projects, 28,763 first-party (26,948 direct from ISO/RTO filings, 1,815 from state programs); the remainder reach us only via LBNL's annual file or a competitor's product. Live per-market breakdown at /catalog. For investors, developers, and AI agents.

Aggregates are keyless Rows need a free key Seven ISO queues, daily

Try it now, without signing up

GET /stats · no credential

This runs against production from your browser — the same host your code will call, the same rows. No sandbox, no fixture, no key.

bash
curl https://api.prospectorlabs.io/stats

See what's actually in there

computed live, per market

The question right after the sample query is always the same one. Per-market totals, how much of each market we pull directly from the ISO versus take from an aggregated source, and when each one last refreshed — computed live, including the markets where our coverage is weak.

Coverage by market →

Connect your AI agent

MCP · Claude Desktop, Claude Code, Cursor

The keyless aggregates (/stats, /catalog/data, /coverage and /source-freshness) work straight away; anything returning per-project rows needs a free key in PROSPECTOR_API_KEY.

claude_desktop_config.json
{
  "mcpServers": {
    "prospector-energy": {
      "command": "uvx",
      "args": ["prospector-energy-mcp"],
      "env": { "PROSPECTOR_API_URL": "https://api.prospectorlabs.io" }
    }
  }
}

Then ask your agent: "What's in the CAISO interconnection queue?"

Get a key for the rows

The aggregates stay open. The rows take thirty seconds.

The live query above ran with no credential, and the aggregates always will. Everything that returns per-project records — the queue itself, the change feed, developers, data centers — needs a key; bulk CSV export needs a paid plan. No email to a human.

API reference Coverage Pricing MCP on PyPI