Contractor licensing complexity across 118 jurisdictions (2026 open dataset)
A machine-readable map of the world's contractor-licensing boards. 118 jurisdictions. 21 EASY. 56 MEDIUM. 24 HARD. 17 BLOCKED. Free to redistribute with attribution.
- Only 21 of 118 contractor-licensing jurisdictions expose direct machine-readable data (Socrata, REST APIs, or bulk JSON). The other 97 require HTML scraping, headless-browser automation, or human operators.
- Four US states — Kansas, Maine, Missouri, New Hampshire — have no state-level general-contractor license. Eight Atlantic Canadian provinces and territories have no provincial residential-builder registry at all. Houston, Texas issues no GC license at any level.
- Quebec's Régie du bâtiment du Québec publishes the gold-standard reference dataset: a daily JSON extract under CC-BY-4.0, roughly 25–30K active records, commercial reuse explicitly permitted with attribution. It is the best contractor-licensing open-data surface on the North American continent.
- Latin American jurisdictions (Mexico, Brazil, Argentina, Colombia, Peru) license the engineer-of-record, not the construction company. AskBaily's matcher models this explicitly — a Bogotá remodel matches a Copnia-registered engineer, not a company license.
- AskBaily has automated 6 of the 21 EASY-tier jurisdictions (California, Oregon, Washington, NYC HIC, Quebec, Indiana) as of Wave 181. The remaining 15 EASY-tier targets are on the Q2 2026 roadmap.
Tier distribution by region
Each jurisdiction was surveyed by identifying the canonical licensing authority, probing its public lookup surface, and classifying the programmatic-access cost into four tiers. EASY means direct machine-readable data. MEDIUM means stable HTML form scraping. HARD means headless-browser automation required. BLOCKED means no online lookup exists or no licensing regime applies.
| Region | Total | EASY | MEDIUM | HARD | BLOCKED |
|---|---|---|---|---|---|
| United States | 60 | 13 | 34 | 8 | 5 |
| Canada | 13 | 1 | 3 | 1 | 8 |
| United Kingdom | 10 | 1 | 4 | 5 | 0 |
| Ireland | 3 | 0 | 3 | 0 | 0 |
| Australia | 8 | 1 | 7 | 0 | 0 |
| New Zealand | 1 | 1 | 0 | 0 | 0 |
| APAC | 9 | 2 | 2 | 3 | 2 |
| MENA | 8 | 0 | 0 | 4 | 4 |
| LATAM | 6 | 0 | 2 | 4 | 0 |
| Total | 118 | 19 | 55 | 25 | 19 |
What the tiers actually mean
EASY (21 jurisdictions)means the licensing authority publishes direct machine-readable data. Oregon's Construction Contractors Board exposes Socrata dataset g77e-6bhs with 21 fields per license, including bond and insurance amounts. Washington's L&I publishes dataset m8qx-ubtq with 20-plus fields. Quebec publishes a daily JSON extract under CC-BY-4.0 — the only licensing authority in North America that has licensed its own data for commercial reuse with attribution. New York City's Department of Consumer and Worker Protection publishes dataset acd4-wkax covering all five boroughs' Home Improvement Contractors. Indiana's Professional Licensing Agency exposes a credentialed MuleSoft REST API with real-time writes.
MEDIUM (56 jurisdictions)means a stable HTML form or ASP.NET page. California's CSLB is MEDIUM despite being the flagship US state — the verification portal is server-rendered HTML with no JavaScript gate, and CSLB explicitly permits programmatic extraction. Nevada's NSCB exposes a querystring shortcut that bypasses the usual POST round-trip. Most K-N states (Kentucky, Louisiana, Maryland, Michigan, Minnesota, Nebraska, North Carolina, New Mexico) fit this tier.
HARD (24 jurisdictions)means headless-browser automation is required. Arizona's Registrar of Contractors runs on Salesforce Experience Cloud; datacenter IPs trigger Cloudflare bot protection, so residential-IP proxy rotation is mandatory. Hawaii's MyPVL migrated to a SPA in 2023 and sits behind a WAF. Massachusetts's dual-registry CSL + HIC flow requires ViewState token handling across two separate ASP.NET portals. Saudi Arabia's Balady portal is Arabic-language with captcha. Brazil's CREA system is federated across 27 state councils.
BLOCKED (17 jurisdictions)means no public online lookup exists or no licensing regime applies. Kansas, Maine, Missouri, and New Hampshire do not license general contractors at the state level. Manitoba, Nova Scotia, New Brunswick, Newfoundland & Labrador, Prince Edward Island, Yukon, Northwest Territories, and Nunavut have no provincial residential-builder registry. Qatar, Kuwait, Oman, Bahrain, and Indonesia have no public contractor-verification surface. For these, AskBaily's operator queue performs a 72-hour manual verification against the municipality of record or the registrant's own filing history.
Citation
If you cite this dataset, please reference:
AskBaily. (2026). Contractor License Verifier Coverage Dataset.
https://askbaily.com/data/license-verifier-coverage.json
Corrections to [email protected] with subject “LVC dataset correction” and the jurisdiction ID. Dataset licensed under CC-BY-4.0. Version 1.0.0, published 2026-04-23.
Frequently asked questions
What is the AskBaily Contractor License Verifier Coverage dataset?
It is a machine-readable CC-BY-4.0 dataset documenting 118 contractor-licensing jurisdictions across the US, Canada, UK, Ireland, Australia, New Zealand, APAC, MENA, and LATAM. Each jurisdiction entry includes the licensing authority, public lookup URL, API type (REST, Socrata open-data, HTML scrape, Playwright proxy, or manual), a four-level complexity tier (EASY / MEDIUM / HARD / BLOCKED), AskBaily's automated-verify status, license-number format regex where available, and extractable fields. The companion JSON feed is at /data/license-verifier-coverage.json.
Why did AskBaily publish this as CC-BY-4.0 open data?
Because the current state of contractor-license verification is a research problem, not a competitive moat. Angi, Thumbtack, and HomeAdvisor treat contractor directories as proprietary because their revenue model is selling lead access. AskBaily's revenue model is a closed-job take-rate, which means the underlying regulatory-access data has no conflict with open publication. Journalists, academic researchers, competing platforms, and government transparency projects can redistribute or republish this dataset with attribution.
What does the EASY / MEDIUM / HARD / BLOCKED tiering mean?
EASY means the licensing authority publishes direct machine-readable data — a Socrata open-data endpoint, a REST API (credentialed or not), a bulk CSV download, or a daily JSON dump. 21 of 118 jurisdictions qualify. MEDIUM means a stable HTML form or ASP.NET page that scrapes server-side without JavaScript execution — 56 of 118. HARD means a JavaScript-rendered SPA (Salesforce Experience Cloud, React, Power Apps), an aggressively WAF-protected portal, or CAPTCHA gating — 24 of 118, requiring a headless-browser microservice. BLOCKED means no online lookup exists at all, or no licensing regime exists for the relevant work — 17 of 118, requiring manual ops intake.
Which jurisdictions are BLOCKED, and why?
Seventeen jurisdictions are BLOCKED. They break into three groups. First: four US states with no state-level general-contractor license (Kansas, Maine, Missouri, New Hampshire — GC licensing is municipal or absent). Second: eight Canadian provinces and territories with no residential-builder registry at all (Manitoba, Nova Scotia, New Brunswick, Newfoundland and Labrador, Prince Edward Island, Yukon, Northwest Territories, Nunavut). Third: five jurisdictions with no public contractor-verification surface (Indonesia, India, Qatar, Kuwait, Oman, Bahrain — state-level registries either require in-country credentials or publish nothing machine-readable). For these, AskBaily routes to a human operator with a 72-hour SLA.
Which jurisdictions has AskBaily already automated?
Six as of 2026-04-23: California (CSLB), Oregon (CCB via Socrata g77e-6bhs), Washington (L&I via Socrata m8qx-ubtq), New York City — Home Improvement Contractor (NYC DCWP via Socrata acd4-wkax), Quebec (Régie du bâtiment du Québec via CC-BY-4.0 open-data JSON), and Indiana (PLA MuleSoft REST API). All six shipped in Wave 181 of AskBaily's pro-onboarding pipeline. The remaining EASY-tier jurisdictions (22 total) are targeted for rollout by end of Q2 2026.
What is the single most surprising finding from this research?
Houston has no general-contractor license at all. Any contractor claiming a Houston GC license is misrepresenting. This is not a gap in the research — it is a regulatory choice by the City of Houston and the State of Texas. AskBaily's Houston matcher verifies TDLR trade cards (electrician, HVAC, plumbing) plus a Texas Secretary of State business-entity check. Dallas and Austin, by contrast, both run Accela Citizen Access portals that AskBaily verifies through a shared adapter. This Texas four-layer composite (state TDLR + Houston no-license + Dallas Accela + Austin AB+C) is the single most complex US regulatory topology.
How does AskBaily handle licensing regimes that license the engineer rather than the company?
Latin American and parts of the MENA region license the engineer-of-record, not the construction company. Colombia (Copnia), Peru (CIP), Argentina (CPIC), Mexico (SEP Cedula), and Brazil (CREA + CAU) all use this model. AskBaily captures this explicitly in the dataset's apiType and notes fields and in the matcher logic — a residential remodel in Bogotá is matched against a Copnia-registered engineer, not a company license. This is a deliberate modeling choice: pretending these jurisdictions license companies would produce false positives and mislead homeowners.
How should academic papers, journalists, and policy researchers cite this dataset?
Use this citation string: 'AskBaily. (2026). Contractor License Verifier Coverage Dataset. https://askbaily.com/data/license-verifier-coverage.json' and attribute 'Source: AskBaily (askbaily.com), CC-BY-4.0.' For corrections or updates to a specific jurisdiction's tier or board metadata, email [email protected] with the subject 'LVC dataset correction' and the jurisdiction ID. The dataset is versioned; version 1.0.0 (2026-04-23) is the first public release, with monthly refresh cadence thereafter.