Most fake address data pairs a Brooklyn street with a Bronx ZIP code. This one picks the neighborhood first, so the street, the ZIP, and the city line agree.
Fake address data usually breaks in the same place. The street is real, the ZIP code is real, and the two have nothing to do with each other. Anyone who knows the city spots 4210 Bedford Avenue, Bronx, NY 10021 in a second, because Bedford Avenue runs through Brooklyn and 10021 sits on the Upper East Side. Test data with that kind of seam in it makes a demo look careless and makes a QA ticket harder to read.
The draw starts with a neighborhood, not with a street. Once Flushing or Bay Ridge or Mott Haven is chosen, the street comes from a list of roads that run through it, the ZIP code comes from the set the Postal Service assigns there, and the city line follows the local convention. Nothing crosses over between boroughs.
| Borough | City line on the label | ZIP range | House numbering |
|---|---|---|---|
| Manhattan | New York | 10001 to 10282 | Plain numbers, tied to the grid |
| Brooklyn | Brooklyn | 11201 to 11256 | Plain numbers, restarting per street |
| Queens | Astoria, Flushing, Jamaica and 50 more | 11101 to 11697 | Hyphenated, cross street first |
| The Bronx | Bronx | 10451 to 10475 | Plain numbers, rising north |
| Staten Island | Staten Island | 10301 to 10314 | Plain numbers, rising south |
| Westchester and upstate | Yonkers, Albany, Buffalo, Rochester, Syracuse | 10601 to 14222 | Plain numbers |
Mail to Queens carries the neighborhood name in the city line. A package to Bell Boulevard goes to Bayside, NY 11361, and one to Austin Street goes to Forest Hills, NY 11375. Writing Queens there is one of the fastest ways to mark a dataset as invented, so the neighborhood name is on by default and the checkbox turns it off if your form validation was built around borough names.
An address like 34-21 30th Avenue reads as two parts. The 34 is the nearest cross street, in this case 34th Street, and the 21 is the position along the block. The scheme dates to the 1920s Philadelphia plan, adopted so that a borough absorbing dozens of separate villages, each with its own Main Street and its own numbering, could be walked by someone who had never been there.
The generator keeps that promise. The number before the hyphen is drawn from the cross street range that actually runs through the neighborhood, so a Ridgewood address lands in the fifties and sixties while a Jamaica address lands in the hundreds. Turning the hyphen off collapses the number into the run-together form, 3421 30th Avenue, which some systems store and print instead.
Above 14th Street, a building number on an avenue tells you which cross street it sits near. Drop the last digit, halve what remains, then add the key for that avenue.
Keys are fixed per avenue: 3 for First and Second, 10 for Third and Eighth, 13 for Ninth, 14 for Tenth, 22 for Lexington, 26 for Madison, 35 for Park, 60 for Amsterdam, Columbus, and West End, and minus 30 for Broadway above 23rd Street. Central Park West and Riverside Drive break the pattern and divide by ten instead of two. Fifth Avenue has five separate keys depending on how far uptown the number falls.
This tool runs the arithmetic backwards. It picks a cross street inside the neighborhood first, then solves for the house number that would land there, which is why a Chelsea address on Tenth Avenue comes back in the low 200s and an Upper West Side address on Amsterdam comes back in the 500s. Turn on the cross street hint and each card prints the street the number resolves to, so the output doubles as a way to check the formula by hand.
What these addresses are not
They are plausible, not deliverable. Nothing here is checked against the USPS delivery point database, so a generated house number may sit past the end of the block, land on a lot that is a parking garage, or duplicate a real apartment where someone lives. Street to ZIP pairing is right at the neighborhood level and not at the block level, so a long avenue crossing three ZIP codes may come back with a ZIP from the wrong end of it. The Queens hyphen has genuine exceptions, including Broad Channel and the named streets in Forest Hills Gardens, which the generator applies the hyphen to anyway. If you need addresses that pass CASS or DPV validation, this is the wrong tool, and a postal validation service is the right one.
Every result comes from a deterministic stream keyed to the seed shown above the cards. Type SPRINT-42 into the seed field and the same six addresses come back tomorrow, on a colleague's laptop, and inside your CI run. Change the area or a checkbox and the same seed produces different output, since the settings feed the draw. The URL accepts ?seed=SPRINT-42, so a link reproduces a fixture set without a file.
Copy all gives one address per line in the single line form, ready for a spreadsheet column or a text fixture. Copy JSON gives an array of objects with the parts kept apart, so street, unit, city, state, zip, borough, neighborhood, and type each land in their own field. Empty units come through as an empty string rather than being dropped, which keeps the shape stable for a parser. Download CSV writes the same columns to a file named after the seed.
For the rest of a test record, the random ZIP code generator covers other states, and the random phone number generator produces the matching contact line.
Generation runs in your browser. No address is sent anywhere, nothing is stored after you close the tab, and the seed is the only thing worth keeping. Output is free to use in test suites, demos, published fiction, and course material with no attribution.
One line worth holding: an address paired with a real name is a claim about where a person lives. Do not use generated addresses to open accounts, pass identity checks, register a business, file paperwork, or make a listing look local when it is not. Those are fraud, whatever the data came from.
Accuracy, ZIP pairing, Queens numbering, and where the output holds up.
The street names, neighborhoods, and ZIP codes are real. The house numbers are drawn at random inside plausible ranges, so a result sometimes matches a building that exists and sometimes points at a stretch of road with no building on it. Treat every result as invented and never send mail to one.
At the neighborhood level, yes. Each neighborhood carries the ZIP codes the Postal Service assigns to it, so a Park Slope address comes back as 11215 or 11217 and never as a Bronx code. Block level accuracy is a different matter. Flatbush Avenue runs through six ZIP codes, and the generator picks one from the neighborhood pool rather than working out which stretch the house number sits on.
Because that is how mail to Queens is addressed. The Postal Service uses neighborhood names as the city line for Queens, so Bayside, Flushing, Jamaica, and Forest Hills each act as their own mailing city. Manhattan uses New York, and the other three boroughs use their own names. Turn off the neighborhood checkbox if your validation rules were written around the word Queens.
The part before the hyphen is the nearest cross street and the part after it is the position along the block, so 34-21 30th Avenue sits on 30th Avenue just past 34th Street. Queens adopted the scheme in the 1920s while merging villages that each had their own street numbering. Some systems store it without the hyphen, which the checkbox handles.
Format checks and ZIP to state checks pass. Anything backed by the USPS delivery point database fails, because these house numbers were never verified against a real delivery point. That difference is often what you want in a test, since it exercises the failure path most address forms never see.
Copy the seed printed above the cards and paste it into the seed field with the same area, building type, count, and checkboxes selected. The output is identical every time, since nothing is stored between visits and the draw is fully determined by the seed. Adding ?seed=yourseed to the page URL loads the same batch straight away.
Yes. The upstate and Westchester option covers Albany, Buffalo, Rochester, Syracuse, Yonkers, and White Plains with local street names and their real ZIP codes. The statewide option mixes those in with the five boroughs at roughly the ratio the state population sits at, so most results still land in the city.
For software testing, demos, screenshots, teaching, and fiction, yes. Using one to open an account, pass an identity check, claim residency, or register a business is fraud in every state, and a generated address is no safer than an invented one.