ordinexPoint-in-time energy intelligence
Energy signals built for
systematic trading.
EIA, NOAA, ERCOT & ENTSO-E energy-market data, transformed into point-in-time, walk-forward-validated factors - delivered through one production-ready API. Not a data reseller: primary-source ingestion, reporting-lag safety, and out-of-sample validation, end to end.
Built for systematic traders · quant researchers · trading bots · energy-focused funds · financial data developers
Not another data reseller
What makes Ordinex different?
A typical energy API stops at raw data - you still do the point-in-time correction and factor research yourself. Ordinex does that work upstream: primary source → PIT normalization → validation → factor research → API → your trading system.
| Capability | Typical energy API | Ordinex |
|---|---|---|
| Point-in-time correctness | Rare | ✓ |
| Reporting-lag aware | Rare | ✓ |
| Walk-forward out-of-sample tested | No | ✓ |
| Economic significance scored | No | ✓ |
| Raw + derived (factor) data | Raw only | ✓ |
| REST API | ✓ | ✓ |
| Reseller / third-party aggregator | Often | No |
Build vs. buy
Why not just pull EIA, NOAA, ERCOT, and ENTSO-E yourself?
All four sources are free and public. Plenty of teams do pull them directly - here's what that actually takes, beyond the download.
Why this matters
No look-ahead bias, and no accidentally-honest factor
In plain terms: we test whether a signal actually survives out-of-sample before it ever reaches the API. Here's how that discipline is enforced end to end:
Most energy-data feeds stop at "here's the number." This one goes one step further: every factor was tested for whether it actually predicts anything, walk-forward, out-of-sample - and the honest misses are published right alongside the real ones. Full breakdown: the validation methodology and the factor catalogue.
Source publishes
EIA reports a weekly storage figure days after the "as of" date it actually describes.
Reporting lag respected
A real min_lag_days safety margin per source - a factor is never paired against a target move that happened before it was actually public.
Nothing overwritten
Every raw fact and factor observation is append-only, insert-if-not-exists - a later revision never silently replaces history.
Walk-forward tested
One held-out fold per calendar year, never fit and tested on the same window.
Scored, not just shipped
A five-dimension quality scorecard decides the tier - sign consistency, crisis robustness, and economic effect size, not just "it correlates."
Architecture
From raw energy data to trading-ready signals
Primary sources
EIA v2 (natural gas, storage, petroleum, electricity) and NOAA CDO, direct public-domain federal sources, plus ERCOT's own Public API for real-time settlement-point power pricing. CAISO/NYISO wholesale power still feeds the internal factor library only.
Point-in-time normalization
Append-only storage - historical queries are reconstructed using each source's own recorded publication timing and reporting-lag rules, refreshed automatically (daily or weekly, matched to how often each source actually updates).
Factor construction
Storage/weather/spread-based signals computed forward from the normalized data - the raw material every factor below is built from.
Walk-forward validation
Bucket analysis, regime splits, annual walk-forward, and an economic-significance effect size - before a factor ever reaches the API.
Production API
Raw series, factor observations, rankings, and screens - all over HTTP, scoped to what's actually legally cleared to serve.
Your trading system
One X-API-Key header away from a live factor feed - built to be called from code, not read from a dashboard.
Factor library
Tiered by real evidence, not an undifferentiated list
A deliberately small factor library. We don't ship hundreds of weak signals - only factors that survive our validation pipeline reach production.
Every factor below is derived from EIA, NOAA, ERCOT, or ENTSO-E data - sources our own legal review has covered for this API's commercial use (see the legal review). Factors are tiered by how they scored on a five-step validation framework: predictive content, monotonicity, regime stability, mandatory walk-forward out-of-sample testing, and economic significance - not just whether they "worked" once. Backtested effect sizes below are historical and gross of transaction costs, slippage, and capacity - not a forward return guarantee. The full methodology and every real backtest result lives in the factor catalogue.
Full breakdown of every factor's own validation evidence and economic effect →
Pricing
Free during early access - pricing not final
Early access is genuinely free, not a bait-and-switch trial - the tiers below are the honest direction pricing is headed, not a live checkout.
Early Access
- Full public factor library
- Raw EIA series (
/v1/series) - Standard rate limit
Pro
- Higher rate limits
- Extended history
- Additional factor coverage
- Commercial usage rights
- Priority support
Enterprise
- Custom rate limits
- Dedicated data requirements
- Commercial licensing
- Research / data partnerships
Quickstart
From zero to your first factor score
Get an API key
Sign up below for a free self-service key - just an email address, no credit card.
Call an endpoint
Send it as the X-API-Key header on every request.
Read the response
The top 5 storage regions ranked by their latest percentile value, most recent observation on/before today.
Reference
Full endpoint reference, generated from the API itself
Every request/response shape below is generated straight from the running API, so it never drifts out of sync with what's actually deployed.
Access
Start building with Ordinex
Free API access during early access. No credit card - start with the full public factor library.
One active key per email. Every self-service key gets the full public factor library above; the CAISO/NYISO spark-spread factors stay internal-only regardless of key, pending the ISO licensing question noted in the factor library.
Prefer to talk first, or need a higher rate limit?