Remodeling contractors: the verified database

How a record gets into the store

Every licence in this store was read from the issuing authority’s public register. The source is a government licence file or a municipal contractor registration — reliability tier 1 — not a Google listing, a review site or a paid directory. Each record stores the licence number, the classification, the status as the board printed it, the authority’s name, a source URL, and the date we retrieved it.

Of 22,183 licences, 21,956 are coded primary_verified_active, 183 inactive, 35 suspended and 9 revoked. Those four statuses sum to the licence total. Suspended and revoked are board findings; they are not a HyreRemodel ranking.

Retrieval dates cluster on three harvest days: 11,900 records on 30 August 2026, 9,721 on 29 August, and 562 on 16 August. The snapshot date (2026-09-05) is when we counted the files, not when every board last changed a row.

The harvest is not the store. Florida’s DBPR files, Arizona’s ROC classification sweep, California’s CSLB name cursor and Nevada’s uncapped listing search all returned more rows than we promoted. A Certified General Contractor who builds hospitals and never remodeled a kitchen can sit in a harvest file and never become a company in this store. The 22,183 figure is the licence layer we kept, counted from files on disk, not a claim that we hold every remodeler who advertises in these seven states.

Florida has no single statewide query — a full DBPR harvest iterates county codes. Arizona’s registrar caps a classification search near a thousand rows, so large classes are partitioned by city. North Carolina’s portal rate-limits and stops a query at about 300 rows. California’s board has no classification search at all, so the roster is reachable only by name. Utah’s lookup has no class filter and sits behind reCAPTCHA. None of those constraints is a reason to pad the count; they are reasons the North Carolina cell is small and labelled incomplete, and reasons the California and Utah cells are lower bounds.

A trade licence is authorisation. It is service evidence only when the class name says so. Arizona’s B-3 (General Remodeling and Repair) and the CR-61 / C-61 / R-61 family name the trade. Nevada’s B-7 does too, and it is young and small. Florida’s CGC, CBC and CRC do not. California’s Class B does not. Utah’s B-100 does not. North Carolina’s Building and Residential classes do not. Most of this store is the second kind: the firm is licensed to do the work, and whether it sells remodeling is a separate claim.

The class-inclusion problem: authorised to remodel is not “is a remodeler”

Class inclusion is the defect remodeling is most prone to, and it is why a board harvest is not a remodeling directory. A trade licence proves authorisation. Unless the class name names remodeling, it is not service evidence. Roofing was unusually clean — Florida CCC, California C-39, Arizona CR-42 all say “roofing” in the class name, so the licence itself was service evidence. Remodeling is the opposite.

Florida licenses remodelers as CGC, CBC or CRC. North Carolina licenses them as Building or Residential. California’s anchor is Class B. Utah’s is B-100. Nevada’s bulk is B and B-2. None of those class names says remodeling. A general building licence permits a kitchen remodel in every one of those states and evidences nothing about whether the company sells one. Including every such licence as a “remodeling contractor” would be a category error: it would count hospital builders, production homebuilders and commercial GCs as remodelers because they are allowed to swing a hammer in a dwelling.

Arizona is the exception that writes the trade into the classification. B-3 General Remodeling and Repair Contractor, the CR-61 / C-61 / R-61 Carpentry, Remodeling and Repairs family, and R-62 Minor Home Improvements are service evidence — the registrar wrote the scope down. B, KB-1 and KB-2 in the same state are not. Nevada is a younger, smaller exception: B-7 Residential Remodeling names the trade, and in the 29 August 2026 listing it accounted for 82 holders across both variants, against 5,034 rows in the two families that only authorise. Arizona (B-3 / CR-61 / C-61 / R-61 / R-62) and Nevada (B-7) are the only regimes in this store where the class name itself can carry a service claim.

HyreRemodel analysis: the inclusion rule has to be stated per state, not once for the hub. In Arizona and for Nevada B-7, the board harvest of the named class is a remodeling roster for that class. In Florida, California, Utah, North Carolina and the rest of Nevada, the board harvest is a roster of firms authorised to do the work. A second signal — the firm’s own site, a BBB type-of-business category, an NKBA or NARI membership, a permit record, or, as a weaker company-side signal, a trading name that names the trade — is what turns authorisation into a service claim. Where we have only the class, the honest label is authorisation, and the company-service row stays inferred. That is why this hub publishes counts of licences, not a searchable list of “remodeling contractors near you.”

The same rule is why we will not ship a fake search box on this page. A search that returned every CGC in Miami-Dade as a remodeler would be a working interface built on a false inclusion rule. The board’s own lookup is the search. Ours is the method.

How to read a row in the table above

Take the Florida row. Licences: 7,130. Share of store: 32.1%. What we hold: a state licence. Issuing authority: Florida DBPR, Construction Industry Licensing Board. Coverage: statewide board harvest; class names do not say remodeling. Read it as: we retrieved 7,130 CILB credentials (CGC, CBC, CRC and related) from the public Verify a Licensee form, on 30 August 2026, by iterating county codes, and we kept those records in this store. Do not read it as: Florida has 7,130 remodeling companies. One company can hold more than one licence. A CRC in the harvest who never advertised a remodel can still be in the file. A remodeler who holds only a local licence in a municipality that also requires one will not appear unless they also hold the state credential we queried.

Take the Texas row. Licences: 955. What we hold: municipal registration. Issuing authority: City of San Antonio (805) and City of El Paso (150). Coverage: two cities only; TDLR licenses no residential general contractor; not a Texas roster. Read it as: we retrieved two city registers, on 29–30 August 2026, and we stored 955 registration files. Do not add 955 to 7,130 and call the sum “verified remodelers.” The instruments are not the same. Do not infer that Houston, Dallas or Austin have no remodelers; those cities are outside this harvest. Do not treat a San Antonio “City Residential Building Contractor” number as a TDLR licence. The 150 company-state rows behind the Texas cell are all primary_verified_alternate and all El Paso; San Antonio sits in the licence layer without a matching Texas company-state count of 805.

Take the North Carolina row. Licences: 149. What we hold: state licence (partial harvest). Coverage: 149 records; the board rate-limits; G.S. 87-1 leaves work under $40,000 unlicensed at state level. Read it as: we hold a capped set. Fourteen city queries came back at the 300-row ceiling. The cities we queried were launch metros, not Wilmington, Asheville or Greensboro. A $35,000 kitchen in Raleigh is lawfully outside this board. Publishing 149 as “North Carolina remodelers” would be a fabrication. Publishing it as a partial harvest with those three limits attached is the row.

The share-of-store column is a concentration warning, not a market-share ranking. Florida, Arizona and California together hold 16,882 licences — 76.1% of the store. That is where we have been able to harvest at scale, not a finding that those states have more remodelers per household. Utah at 703 is a name-sweep lower bound. North Carolina at 149 is a capped set. Comparing those two percentages as if they were the same kind of number is the mistake the coverage column exists to stop.

Bars under the table are the same counts. Longer is not better. A longer bar on Florida means a larger retrieved set of CILB credentials that survived inclusion, not a recommendation and not a claim that Florida’s board is stricter. A short bar on North Carolina is a harvest limit plus a statute, and reading it as “few remodelers” would invert both.

How this hub relates to the census and the lookup directory

Three pages share this store and split the work so that none of them has to pretend to be the other two. This hub is the public face: the per-state counts, the verification method, the harvest routes and the coverage limits, stated before anything that looks like a directory. It answers “what does HyreRemodel actually hold, and what must I not infer from it?”

The remodeling contractor census is the method study of the same files. It owns inclusion rules, the difference between a harvest file and a promoted company, the reason a board harvest over-counts authorised firms and under-counts firms whose class does not exist, and the measured 2026-09-05 snapshot as a research object. If this hub is the table, the census is the codebook. Counts on both pages must match; if they ever do not, the snapshot was re-measured and one of the pages has not yet been updated.

The contractor licence lookup directory is the instrument a homeowner actually uses. It names the issuing board for each state we cover and the public URL a permit office will recognise. It does not re-count the store. It does not rank firms. It exists so that step 4 above has a durable address, and so that a reader who arrived here looking for a search box leaves with a working one — the board’s.

Three more pages sit next to those two. What the licence class actually names is the classification study: Florida, California, Nevada, Utah and North Carolina mostly license a building class; Arizona (and Nevada B-7) write the trade into the class. States without contractor licensing owns the statutory finding that Texas — and other verified no-licence states — issue no residential GC credential; this hub only carries the two municipal registers we have harvested. Texas: two cities, no state licence is the market page for that finding, with the San Antonio and El Paso rules, fees and live records.

HyreRemodel analysis: splitting the store across a hub, a census and a lookup is the alternative to a fake search product. A single URL that tried to be all three would bury the caveats under a directory, which is how every scraped contractor site on this query goes wrong. The census can go as deep as the files allow. The lookup can stay current as boards move their URLs. This hub can stay honest about what the 22,183 figure is. None of them is a 50-state roster, and none of them will become one by adding Texas municipal rows until the rest of Texas is harvested.

How to check a contractor yourself

The store is a research layer. The authority of record is still the board. If you have a licence number — or a business name and a state — use the issuing register. That check is free, it is current as of the board’s own clock, and it is the check a permit office will run. Our retrieval dates are 16, 29 and 30 August 2026; the board’s clock is today.

Florida DBPR — Verify a Licensee · Arizona ROC — Contractor Search · California CSLB — Check a License · Nevada NSCB — License Search · Utah DOPL — Licensee Lookup · North Carolina NCLBGC — Public Search · San Antonio Contractor Connect · El Paso Accela Citizen Access.

For Texas work outside those two cities, there is no state general-contractor number to check. Confirm the city’s own registration if it has one, then the trade licences (electrical, plumbing, air-conditioning) that TDLR does issue, then a certificate of insurance from the insurer. Absence of a Texas GC licence is the legal baseline, not a red flag.

A licence is a floor. Comparing what each quote actually includes — drawings, allowances, exclusions, change-order rules, payment schedule, lead-safe certification if the house is pre-1978 — is how you tell a complete bid from a cheap one.