Developers · API & MCP

Build with PowerGridIQ: grid intelligence your agents can call.

As compute goes agentic, the question "where and when should this electricity load run" gets answered by machines, continuously. PowerGridIQ gives your agent a reasoned, cited decision, not a raw feed: the PGIQ Rating across 84 global power markets, as a JSON API and a Model Context Protocol (MCP) server. Free to evaluate, no key.

What it is, and why an agent wants it

Raw grid data, carbon intensity, price, is a commodity an agent still has to reason over. PowerGridIQ sells the reasoning. Every market is scored on five pillars (access, availability, cost, momentum, carbon), assigned a tier (1 Prime to 5 Largely closed) and an outlook, and backed by evidence and sources. The /best endpoint returns a ranked answer under the lens you choose, so your agent asks a question and gets a decision with a rationale and a confidence level.

Every market carries a confidence field and a provenance note. The rating is a directional screen, not a connection guarantee, the interconnection queue is the real gate. Build that caveat into anything that acts on the output.

Quickstart

Base URL https://powergridiq.com/api/v1 · JSON · CORS-open · no key for the evaluation tier.

Ask for a decision
# Best carbon-light markets in Europe, top 3
curl "https://powergridiq.com/api/v1/best?lens=carbon&group=europe&limit=3"
{
  "lens": "carbon", "group": "europe", "count": 3,
  "results": [
    { "id": "sweden", "tier": 2, "tier_name": "Strong", "lens_score": 83,
      "outlook": "Stable", "confidence": "High",
      "why": "Cheap, clean, open, an established Nordic hub...",
      "report_url": "https://powergridiq.com/rating/sweden" },
    { "id": "norway", ... },
    { "id": "france", ... }
  ]
}
Read one market in full
curl "https://powergridiq.com/api/v1/ratings/quebec"
{
  "id": "quebec", "tier": 2, "score": 74, "outlook": "Positive", "confidence": "Med-High",
  "pillars": { "access": 55, "availability": 90, "cost": 100, "momentum": 30, "carbon": 100 },
  "pillar_rationale": { "access": "Managed access: Bill 69 requires sign-off over 5 MW...", ... },
  "latest_action": { "type": "Affirmed", "trigger": "Hydro-Quebec tariff proposal (Jun 2026)" },
  "sources": [ { "label": "Hydro-Quebec rate, CBC", "url": "..." } ]
}

Endpoints

EndpointReturns
GET /api/v1/metaMethodology, tier legend, pillar weights, the weighting lenses, and the provenance note. Read first.
GET /api/v1/marketsCompact list of all rated markets: id, tier, score, outlook, under-review flag.
GET /api/v1/ratings/{market}Full rating for one market: five pillars with rationale, thesis, peers, sources, latest action.
GET /api/v1/ratingsAll ratings in one call. The single-market fields are free; the compiled fields (realized-cost bands and precise queue waits across every market at once) require a paid key. Fetch any one market in full, free, at /api/v1/ratings/{market}.
GET /api/v1/bestThe decision. Params: lens (default/cost/carbon/momentum), group, min_tier, limit. Returns a ranked answer with a one-line why.
GET /api/v1/screenThe constraint screen. Params: max_cost ($/MWh), max_months, min_tier, min_carbon (0-100), group. Returns every market meeting all your constraints, best tier first. The qualifying set, tier and outlook are free; the precise cost midpoints and wait months need a paid key.
GET /api/v1/demand-signalThe buyer-demand signal: where large-load buyers are actively tracking markets on PowerGridIQ, as a normalized 0-100 interest index and a High/Moderate/Low band per market. Aggregate and privacy-safe (small markets omitted, no personal data). Free and unmetered.
GET /api/v1/gridCompact grid metrics for every market: price, carbon intensity, renewable share, system firm margin, peak stress. Live where a feed exists, else a labelled modelled baseline.
GET /api/v1/grid/{market}Grid snapshot: carbon intensity, price, fuel mix, demand, system firm margin and peak stress. The orchestration inputs.
GET /api/v1/cheapest-windowThe cheapest hours to run a flexible load, per market, from the modelled daily price shape. Pass ?market= for one, or omit for the cheapest grids ranked.
GET /api/v1/developmentsDated, cited grid developments. Optional ?market=.
GET /api/v1/actionsRating-action history (initiations, affirmations, under-review). Optional ?market=.
GET /api/v1/usageYour API key's current-month usage and remaining quota. Send your key as a bearer token.
POST /api/v1/subscribeSubscribe an https endpoint to rating-change webhooks (see below). Body {url, events}.

Full machine-readable spec: openapi.json · agent map: llms.txt · manifest: /.well-known/ai-plugin.json

Open headline dataset (free, CC BY 4.0). The PGIQ headline ratings for all rated markets (tier, PGIQ Rating, Fundamentals score, outlook, confidence and overall reliability level) are published as a downloadable dataset you can redistribute with attribution: on GitHub (CSV and JSON) and as live JSON at /data/pgiq-ratings.json. The five pillar sub-scores, detailed reliability reads, point-in-time history and interconnection-queue detail are proprietary: available through the API for evaluation and under a commercial license for production use. Attribute as "PowerGridIQ Rating."
Reading the grid data. The source field tells you the basis per market: a live feed (eia-live, entsoe-live, and so on) is a current instantaneous reading; modelled-baseline is a typical value. Do not compare a live market and a modelled market as if they were the same instant. renewable_pct counts wind, solar, hydro, geothermal and biofuel only, so a clean nuclear grid can show a low renewable share next to a low carbon intensity. firm_headroom_mw is firm capacity minus the annual peak: a negative value (and peak_stress_pct above 100) means peak demand exceeds firm capacity, so the grid leans on non-firm supply and imports at peak. These are structural siting signals, not a real-time adequacy alarm.

Auth, limits, and webhooks

Free to evaluate. Add a key for production volume. Subscribe to changes so your agent re-decides.

Free tier: no key, 100 requests/day per IP, every endpoint above. Keyed tiers: send Authorization: Bearer pk_live_... for a higher monthly quota (developer 25,000/mo, pro 500,000/mo). Every response carries X-PGIQ-Tier, X-PGIQ-RateLimit-Limit, and X-PGIQ-RateLimit-Remaining; over quota returns 429. Check your own usage at /api/v1/usage.

# keyed request
curl "https://powergridiq.com/api/v1/best?lens=cost" \
  -H "Authorization: Bearer pk_live_..."
Webhooks: re-decide when a rating changes

Subscribe an https endpoint and get a signed POST when a market gets a new rating action, enters Under review, or clears. Each delivery is signed with X-PGIQ-Signature: sha256=HMAC-SHA256(secret, body), verify it before acting.

curl -X POST "https://powergridiq.com/api/v1/subscribe" \
  -H "Content-Type: application/json" \
  -d '{"url":"https://you/hook","events":["rating.action","market.under_review"]}'

Pick a plan below to get a key instantly, or talk through production limits: hello@powergridiq.com.

API plans

Machine access by request volume, for wiring the rating into an agent or app. Your key is issued and emailed the moment you subscribe. For the full plan lineup, including the Analyst Desk for people who research or decide where load should go, see the plans and pricing page →

Free
$0
no key needed
  • 100 requests/day per IP
  • Every endpoint
  • For evaluation
Just start calling
Developer · for machines
$29/mo
25,000 requests/month
  • 25,000 requests/month
  • All endpoints + webhooks
  • Key issued instantly
Pro · for machines
$199/mo
500,000 requests/month, high-volume agents
  • 500,000 requests/month
  • All endpoints + webhooks
  • Priority support
Analyst Desk · for people
$600/mo
the working seat for siting and strategy teams
  • The Global Energy Report, the quarterly book
  • Shortlist alerts: told when a market moves
  • Full API and MCP access
  • Rating history, evidence, and PDF briefs
  • The compiled cross-market tables in full
Enterprise
Custom
volume, SLA, data licensing
  • The Connection Friction Feed
  • High volume, SLA, data licensing
  • Founding design partners open
Become a design partner →

Secure checkout by Stripe. Cancel any time from your Stripe receipt. Need a custom volume or an SLA? Talk to us.

MCP server: call it from any agent

Six tools, pgiq_markets, pgiq_rating, pgiq_best, pgiq_grid, pgiq_cheapest_window, pgiq_developments, over the same API. Works in Claude Desktop, Claude Code, Cursor, or any MCP-capable agent.

Install & run
cd pgiq-mcp
npm install
node index.js   # speaks MCP over stdio
Configure in Claude Desktop
{
  "mcpServers": {
    "powergridiq": {
      "command": "node",
      "args": ["/absolute/path/to/pgiq-mcp/index.js"],
      "env": { "PGIQ_KEY": "pk_live_optional_for_higher_limits" }
    }
  }
}

Then ask your agent things like "Use PowerGridIQ to find the best carbon-light market for a 100 MW AI cluster in Europe," or "What's the PGIQ rating for Ireland, and what would move it?" The agent calls the tool, gets the reasoned answer, and cites the market report.

What to build with it

Carbon- and cost-aware compute orchestration. A scheduler that routes training and inference to the cleanest, cheapest, least-stretched grid right now, and avoids the markets that are hard to connect or stretched at peak.

Autonomous siting research. An agent that shortlists markets for a new data center or industrial load, pulls the rating and the evidence, and writes the diligence note with citations.

Energy procurement and carbon-accounting agents. Tools that weigh price against carbon intensity under your own lens and keep a dated record as ratings change.

Trust, limits, and production use

The rating is a proprietary synthesis: pillar evidence is drawn from public operator and regulator sources, some inputs are modelled, and every market is labelled with a confidence level. It is a directional screen for siting and scheduling, not a connection guarantee. The interconnection queue is the real constraint, and we say so in the data.

The free tier is open for evaluation (100 requests/day, no key). Live carbon and price signals and rating-change webhooks are available now; an API key lifts the rate limit for production volume. For an SLA, a data-quality guarantee, and short-horizon forecasts, get in touch and we will set you up.