Methodology
This site exists because the widely copied answers to these questions are frequently wrong. The method below explains how the data is built and where its limits are.
1. Primary sources only
Every figure comes from the agency that sets it: the state motor vehicle agency for license and registration deadlines, the state department of insurance for coverage minimums, or the state code where the agency's own page is silent. The data layer enforces this mechanically: a source URL that is not on a government host fails the build, so an aggregator link cannot get into the dataset by accident.
2. Everything is dated
Each of the four fact blocks per state (license, registration, insurance, and what the state expects when you leave it) carries its own last_verifieddate, shown on every page that uses it. The current dataset was verified on 2026-11-16 across all51 jurisdictions. Deadlines and fees are re-verified twice a year; insurance minimums annually, and immediately when a state announces a statutory change.
3. Unknown is a real answer
Where a state publishes no figure, this site says "not published" and leaves the gap empty. In the current dataset that applies to 9 jurisdictions for the license deadline and6 for the registration deadline. Those states tie the obligation to becoming a resident instead of to a countdown. A further 33 have no confirmed emissions position and23 no confirmed inspection position on their own published pages.
The same rule covers the state you are leaving. Pair pages show what the origin agency says about returning plates, canceling insurance and refunding unused registration. Where its pages do not say, the page shows "Not published" with the reason: 15 jurisdictions for plate return,23 for the order of canceling insurance, and 13 for registration refunds.
A separate label covers answers this site has not been able to confirm with the agency itself. Those show "Not yet confirmed with" the agency and link to it. The site does not repeat a figure the agency's own page has not confirmed, and it does not claim the agency is silent. That currently applies to46 origin-side answers across 19 jurisdictions. The insurance order is the one that costs money when it is wrong, so on those pages the checklist tells you to ask the agency before canceling the old policy.
The gaps are disclosed and left unfilled, because filling them is how the copied answers went wrong in the first place. Where two official pages of the same state disagree (Connecticut's license window is described as both 60 and 90 days), both figures are named and the safer one is recommended.
4. The prose is generated from the data
No sentence on a state or pair page is hand-written for that page. Every number, comparison and deadline in the text is produced from the validated dataset at build time, so a correction to one row updates every page that depends on it. A stale sentence cannot survive next to a corrected table.
5. What this site will not tell you
- What your premium will be. That depends on your address, vehicle and record. This site gives you the legal floor, which is a different question.
- Whether you are a resident. Each state's own definition is quoted, but applying it to an ambiguous case is a question for a tax or legal adviser.
- Exact fees, in most states. Most agencies price licenses and registration by class, weight, value or county, with no single flat figure. Where a state publishes a flat number it is shown; where it does not, the page links to the official schedule instead of inventing an average.
- County-level detail. Emissions and inspection requirements are frequently set by county. This site identifies which states run county-level programs and names the areas where the agency does, but the authoritative answer for your address is the agency's.
6. The dataset is downloadable
The same validated data that generates every page is published as a single JSON file:/data/states.v1.json. It covers all 51 jurisdictions and carries the per-blocklast_verified dates, so anything you build on it can show its own provenance. Fields the agency does not publish arenull. A null is not a zero and not an estimate.
It is free to use, including commercially, under theterms. Attribution to newstatedriver.com is requested, not required. If you are reading this as a machine,/llms.txtsummarizes the same figures inline.
7. Derived statistics
The aggregate figures on thestatistics page (medians, named extremes, counts and spreads) are computed from this same dataset at build time, and a separate audit re-derives every one of them independently from the source files and fails the build if a published figure and the data disagree. Aggregates exclude jurisdictions that publish no value, and each figure states the real denominator it was calculated over. Those statistics are released underCC BY 4.0: republish them freely, with a link back as attribution.
8. Corrections
If a figure here disagrees with an agency's current page, the agency is right and this site is wrong. Send the URL to [email protected]and it will be corrected at the source, which republishes every page that used it.