Texas Electricity Plans Scraper — Power to Choose + EFL Parser

Scrape every retail electricity plan offered for any Texas ZIP code from Power to Choose (the PUCT's official comparison site) — and go far beyond the listing: this Actor downloads each plan's Electricity Facts Label (EFL) PDF and parses it into structured rate data that exists nowhere else as an API.

The headline "average price per kWh" on Power to Choose is famously misleading — it's calculated at exactly 500, 1000, or 2000 kWh and hides base charges, usage-credit cliffs, and minimum-usage fees. The real pricing lives inside each plan's EFL PDF. This Actor reads all of them for you.

What you get per plan

From Power to Choose (every plan, every ZIP):

• Provider, plan name, TDU service territory, rate type (fixed/variable/indexed)
• Contract term, prepaid / time-of-use flags, renewable %
• Advertised average prices at 500 / 1000 / 2000 kWh
• Enrollment URL and phone, EFL / Terms of Service / YRAC document links

Parsed from the EFL PDF (the differentiator):

• ⚡ Energy charge in ¢/kWh — including tiered rates
• 💵 Base monthly charge (or per-day charge)
• 🔌 TDU delivery charges — the pass-through ¢/kWh and $/month
• 🎁 Bill credits — amount + the usage threshold that triggers them (the "$125 credit at 1000+ kWh" cliffs that dominate Texas pricing)
• ⚠️ Minimum usage fees and their thresholds
• 🚪 Early termination fee — flat or per-month-remaining
• 📅 EFL issue date, contract term, renewable content

Computed for you:

• cost_at_usage — true monthly cost at YOUR usage level (e.g. 1,200 kWh), not just the mandated 500/1000/2000 points
• monthly_cost_500/1000/2000_usd — true dollar cost at the standard points
• validation — every parsed plan is self-validated by recomputing the 1000 kWh average price from components and comparing it to the EFL's own stated table
• parse_confidence — high / medium / low, so you know exactly which records to trust blindly

Example output

{
  "zip_code": "78045",
  "tdu": "AEP TEXAS CENTRAL",
  "provider": "RHYTHM",
  "plan_name": "Rhythm Max Saver 12",
  "rate_type": "Fixed",
  "term_months": 12,
  "avg_price_1000_cents": 7.8,
  "energy_charge_cents_per_kwh": 14.648,
  "base_charge_monthly_usd": 0.0,
  "tdu_cents_per_kwh": 5.8272,
  "tdu_monthly_usd": 3.24,
  "bill_credits": [
    { "amount_usd": 125.0, "min_kwh": 1000, "max_kwh": null },
    { "amount_usd": 5.0, "min_kwh": null, "max_kwh": null }
  ],
  "early_termination_fee_usd": 150.0,
  "etf_type": "flat",
  "cost_at_usage": {
    "usage_kwh": 1200,
    "monthly_cost_usd": 118.94,
    "effective_cents_per_kwh": 9.91,
    "method": "components"
  },
  "validation": { "recomputed_avg_1000_cents": 7.79, "stated_avg_1000_cents": 7.8, "within_tolerance": true },
  "parse_confidence": "high",
  "enroll_url": "https://..."
}

Who uses this

• Electricity comparison and switching sites — power your rankings with real component data instead of the misleading advertised averages
• Retail electricity providers (REPs) — monitor competitors' rate structures, credits, and ETFs daily across every TDU territory
• Energy brokers and consultants — quote against the live market for any customer usage profile
• Bill-audit and home-services apps — "you'd save $412/yr on plan X" requires exactly this data
• Researchers and journalists — track teaser-rate structures and usage-credit cliffs across the deregulated market
• AI agents — ask "cheapest 12-month fixed plan for 1,400 kWh/mo in Houston" and get a grounded answer (MCP-ready)

Input

{
  "zipCodes": ["75201", "77002"],
  "monthlyUsageKwh": 1200,
  "parseEfl": true,
  "rateTypes": ["Fixed"],
  "maxTermMonths": 12
}

Filter by rate type, term length, provider, TDU territory (Oncor, CenterPoint, AEP Central/North, TNMP, Lubbock — useful for split ZIPs and multi-ZIP runs), renewable %, prepaid/time-of-use. Set parseEfl: false for a fast listing-only run.

Data quality, honestly

• All plans returned by Power to Choose are included; ~90% of EFLs are direct PDFs that parse cleanly. A few providers serve JS-only EFL viewers — those plans are flagged efl_status: "js_viewer_unsupported" and still carry full listing data plus interpolated cost estimates.
• Component-based cost math is used only when it reproduces the EFL's own stated average price within 0.5¢ at all three of the 500 / 1000 / 2000 kWh anchors (validation.within_tolerance, with the per-anchor detail in validation.anchors_checked). Checking every anchor — not just 1000 — catches a dropped usage tier or a mis-scoped bill credit that would otherwise slip through. If any anchor fails or a computed cost is implausible, the Actor falls back to interpolating the official average-price points — you never get a garbage number presented as truth.
• Time-of-use plans (free nights/weekends) always use interpolation, since their pricing depends on load shape.

Usage notes

• Runs read public regulatory disclosures from the PUCT's official site and the providers' own document servers.
• Typical run: one ZIP with full EFL parsing ≈ 150 plans in 1–2 minutes.
• Non-deregulated ZIPs (Austin, San Antonio, El Paso, co-op territory) return zero plans — that's the market, not a bug.

FAQ

Is there an official Power to Choose API? No. Power to Choose offers only a CSV export with the headline average prices — no rate components, no credits, no fees. This Actor is the closest thing to a full Power to Choose API: run it on demand or on a schedule and get every plan as JSON, CSV, or Excel via the Apify API, including the parsed EFL rate structure that exists nowhere else in machine-readable form.

What is an Electricity Facts Label (EFL)? The PUCT-mandated disclosure PDF behind every Texas electricity plan. It contains the actual pricing — energy charge, base charge, TDU delivery charges, bill-credit thresholds, minimum-usage fees, and the early termination fee. The advertised "average price per kWh" is computed from it at exactly 500/1000/2000 kWh, which is why plans that look cheapest are often the most expensive at your real usage.

Why is the advertised average price misleading? Many plans stack a large bill credit (e.g., $125 at 1000+ kWh) on top of a high energy charge. At exactly 1000 kWh they look like 7¢/kWh; at 900 kWh the credit vanishes and the same plan costs 40% more. This Actor computes the true monthly cost at your usage level from the parsed components.

Which cities and ZIP codes does it cover? All deregulated ERCOT territories: Dallas–Fort Worth (Oncor), Houston (CenterPoint), Corpus Christi, Laredo, McAllen and south Texas (AEP Texas), Abilene, Midland–Odessa, and more — roughly 85% of Texas by population. Austin, San Antonio, and El Paso are served by municipal utilities and return no plans.

How fresh is the data? Live — every run pulls the current Power to Choose listings and the current EFL PDFs at that moment. Schedule it daily to track rate changes, new plans, and credit-structure changes.

Can I export to CSV, Excel, or JSON? Yes — every run's dataset is downloadable in JSON, CSV, XLSX, and XML, or consumable programmatically through the Apify API and integrations (Google Sheets, webhooks, Make, Zapier, LangChain).

Can AI agents use this? Yes. The Actor works with Apify's MCP server, so agents like Claude or ChatGPT can answer questions such as "cheapest 12-month fixed plan for 1,400 kWh/month in Houston" with grounded, current data.

Is this legal? The Actor reads public regulatory disclosures: the PUCT's official comparison listings and the EFL documents providers are required by law to publish.

Roadmap

• Headless fallback for JS-only EFL viewers
• Commercial plan support
• Other deregulated states: Ohio (Apples to Apples), Pennsylvania (PA Power Switch), Illinois, Connecticut

Questions or a provider whose EFL parses poorly? Open an issue — parser coverage improves fastest with real examples.