California Contractor License Scraper — Bond & Workers' Comp

Every one of the 243,555 licensed contractors in California, straight from the official CSLB registry — with the qualification layer that turns a directory into a prospect list.

Other contractor scrapers give you a name and a phone number. This one also tells you whether the contractor can legally work right now, who wrote their bond, when it expires, and whether they carry workers' comp.

Why that matters

Segment	Statewide count	Who buys it
Suspended for no contractor bond	~6,200	Surety agencies — these contractors cannot legally work until they buy a bond
Active with no workers' comp carrier on file	~115,000	Workers' comp brokers
Workers' comp expiring in the next 90 days	~28,000	A dated, recurring renewal trigger
License expiring in the next 90 days	~25,500	Renewal and re-engagement campaigns
Under a disciplinary bond	~1,800	With the reason and case number attached

Each of those is one input toggle away.

Fields

licenseNumber · businessName · businessType · mailingAddress · city · county · zipCode · businessPhone · primaryStatus · secondaryStatus · isActive · classifications (78 codes: B general building, C10 electrical, C36 plumbing, C20 HVAC, C33 painting, C39 roofing, A engineering…) · issueDate · expirationDate · daysUntilLicenseExpiry · pendingSuspension · bondSuretyCompany · bondNumber · bondAmount · bondEffectiveDate · bondCancellationDate · wcCoverageType · wcInsuranceCompany · wcPolicyNumber · wcEffectiveDate · wcExpirationDate · daysUntilWcExpiry · hasWorkersComp · qualifierBond* · disciplinaryBondSurety · disciplinaryBondReason · disciplinaryCaseNo · isUnderDiscipline · sourceUrl · lastUpdate · dataSnapshotDate · dataSnapshotAgeHours · scrapedAt

Measured fill — every field, counted on the whole file

Counted over all 244,204 rows of the live CSLB file (2026-08-19), not a sample. Anything below 100% is below 100% because CSLB does not publish it for that licensee — the absence is the record, and for the workers' comp and bond fields the absence is exactly the sales signal.

Field	Fill	Why it is not 100%
licenseNumber businessName mailingAddress city zipCode businessType primaryStatus classifications issueDate expirationDate daysUntilLicenseExpiry wcCoverageType sourceUrl lastUpdate	100%	—
businessPhone	99.9%	245 licensees list no phone
bondSuretyCompany bondNumber bondAmount bondEffectiveDate	100%	121 licensees exempt from the bond requirement
county	96.5%	out-of-state and PO-box addresses carry no CA county
wcInsuranceCompany wcPolicyNumber	50.1%	no workers' comp carrier on file — the broker lead
wcExpirationDate daysUntilWcExpiry	50.1%	same: no policy, no expiry
fullBusinessName	23.5%	only DBAs and individuals carry a second legal name (57,492 rows: FullBusinessName 41,631 + BUS-NAME-2 21,156, less 5,295 carrying both)
reissueDate	15.7%	only licenses that were reissued
bondCancellationDate	10.1%	only bonds that were cancelled — this is the surety trigger
inactivationDate	8.8%	only licenses that went inactive
qualifierBondSurety qualifierBondNumber qualifierBondAmount	5.6%	13,691 licensees carry a qualifying-individual bond
secondaryStatus	2.4%	only licenses with a secondary condition (probation, pending case)
wcSuspendDate	1.3%	only WC-suspended licenses
wcCancellationDate	1.1%	only cancelled WC policies
disciplinaryBondSurety disciplinaryBondAmount disciplinaryBondReason isUnderDiscipline	0.8%	1,862 licensees are under a disciplinary bond order
disciplinaryCaseNo disciplinaryCaseRegion	0.7%	1,830 of those 1,862 orders carry a case number

isActive, hasWorkersComp, isUnderDiscipline, pendingSuspension, dataSnapshotDate, dataSnapshotAgeHours and scrapedAt are computed, so they are populated on every row.

Example

{
  "businessName": "HYDRO BACKFLOW",
  "city": "WEST COVINA", "county": "Los Angeles",
  "businessPhone": "(844) 554 9376",
  "classifications": ["C36"],
  "primaryStatus": "Contr Bond Susp", "isActive": false,
  "bondSuretyCompany": "NORTH RIVER INSURANCE COMPANY (THE)",
  "bondAmount": 25000, "bondCancellationDate": "2026-07-15",
  "wcExpirationDate": "2026-10-23", "daysUntilWcExpiry": 73,
  "disciplinaryBondReason": "Conditional", "isUnderDiscipline": true
}

A plumbing contractor whose bond was cancelled in July, suspended as a result, under a conditional disciplinary bond, with workers' comp lapsing in 73 days.

Input

input	default	notes
counties	["Los Angeles"]	Empty = whole state. LA 49,637 · San Diego 21,551 · Orange 20,175
classifications	["C36"]	CSLB codes. Empty = all 98. Hyphen optional — C8 and C-8 both match
statuses	—	e.g. CLEAR, Contr Bond Susp, Work Comp Susp
onlyActive	false	Only licenses that are CLEAR
bondSuspendedOnly	false	The surety-lead segment
missingWorkersCompOnly	false	No WC carrier on file
wcExpiringWithinDays	—	Dated workers' comp renewal trigger
licenseExpiringWithinDays	—	License renewal trigger
forceRefresh	false	Bypass the snapshot cache and pull a new copy from CSLB. Needs a run timeout ≥ 15 min
maxItems	5000	Full file is 244,204

How fresh is the data — and how you can tell

CSLB does not publish per-record queries; it publishes one 77 MB bulk file, and it serves that file at about 175 KB/s, so a single download takes 6–8 minutes (measured on the platform: 443 s direct, 344 s through a proxy). Downloading it on every run would make every run cost eight minutes of compute for data that changes a few hundred rows a day.

So this Actor caches the verified file between runs and refreshes it automatically once the cached copy passes 24 hours old. A normal run therefore reads a snapshot instead of re-downloading the state of California.

You are never left guessing how old that snapshot is. Every run tells you three ways:

• the run log: Serving: CSLB snapshot published 2026-08-17 (downloaded …, 3.2 h old, 244,204 licences…)
• the run status message, same line, visible without opening the log
• every row, on dataSnapshotDate (the day CSLB published the snapshot), scrapedAt (when that file was pulled from CSLB — not when the row was emitted) and dataSnapshotAgeHours. dataSnapshotDate is also a column in the default dataset view.

Need this second's data? Set forceRefresh: true and give the run a timeout of at least 15 minutes. The run will pull a brand-new copy from CSLB, re-verify it, emit from it, and leave it in the cache for everyone after you. If the timeout is too short for a 6–8 minute download, the run says so in a warning and serves the cached snapshot rather than failing.

Pricing

$0.004 per license returned ($4 per 1,000). The two other CSLB actors on the store charge $19–21 per 1,000 for license rows without the bond, workers' comp or disciplinary layer.

Every row is charged before it is written, so you are never billed for a row you did not receive and never receive a row you were not billed for. If a run's Max total charge (USD) is lower than maxItems × $0.004, the run stops at that limit, tells you so in the run status, and reports the number of licenses you actually received — it does not quietly hand back a short file.

Reliability note

CSLB publishes this as a single ~77 MB file that carries no content-length and ignores Range, so a truncated response is still syntactically valid CSV and parses without error — a naive scraper silently returns partial state data and reports success. Measured live: one attempt returned 97.3% of the file, the next returned all of it.

This Actor validates every download (trailing newline, complete final record, and the file's legacy tail ending at license 999999) and retries up to four times, refusing to emit anything rather than report a false success. In testing, two consecutive attempts failed before the third came back complete — that is a normal run, not an outage.

The same check is re-applied to the cached snapshot every time it is read, so a corrupted cache can never become silent partial data either. Nothing is ever emitted, and nothing is ever cached, that has not passed the completeness gate.

Data is a public record published by the California Contractors State License Board. You are responsible for complying with CSLB's terms and with how you use the data.