Israel Address Validator (official street list)

Check lists of Israeli addresses against the official locality and street lists published on data.gov.il by the Population and Immigration Authority (רשות האוכלוסין וההגירה). For each address you get the official locality and street names, their official codes (סמל ישוב, סמל רחוב), the house number, and a clear status that says whether the address is valid, needs a human look, or could not be matched.

The lists are downloaded fresh from data.gov.il on every run, so results follow the latest published version.

What it does

• Finds the locality in free text, in Hebrew or English (Haifa, Tel Aviv), and common abbreviations (ת"א, פ"ת, ב"ש, ראשל"צ, ר"ג, ב"ב, י-ם).
• Matches the street against the official street list of that locality only, in this order: exact name → name without street type (רח', שד', דרך, סמטת…) → different word order → spelling variant (ו/י added or dropped) → typo similarity.
• Splits house number, apartment (20/3, דירה 4), floor (קומה), entrance (כניסה) and a 7-digit postal code.
• Never guesses silently: if two streets or localities are equally close, the status is ambiguous_* and you get the candidates as suggestions. Every correction by spelling variant or similarity sets needsReview: true.
• Returns a cleaned address in official spelling: normalizedAddressHe.

Who it is for

E-commerce and delivery teams cleaning checkout addresses, CRM and data teams deduplicating customer lists, and anyone who needs official locality/street codes to join address data with other government datasets.

Input

Field	Type	Notes
addresses	array of strings	Free-text addresses.
records	array of objects	Your own id plus address, or street + houseNumber + city. The id is copied to the output.
defaultCity	string	Used only when an address has no recognisable locality.
onlyProblems	boolean	Output only rows with needsReview: true. The SUMMARY still counts every address.
minStreetSimilarity	number (0.5–1, default 0.8)	Lower fixes more typos but risks wrong matches.
minCitySimilarity	number (0.5–1, default 0.85)	Used only when no locality matches exactly.
maxSuggestions	integer (default 3)	Candidates returned for ambiguous / not-found rows.

Example input:

{
  "addresses": [
    "הרצל 12, תל אביב",
    "שד' רוטשילד 1 ת\"א 6688101",
    "ביאליק 20/3 רמת גן",
    "אבן גבירוך 30 תל אביב"
  ],
  "records": [{ "id": "order-1001", "street": "הרצל", "houseNumber": "5", "city": "Haifa" }]
}

Output

One dataset row per address. Example (real output from a local run):

{
  "id": "addresses[1]",
  "input": "שד' רוטשילד 1 ת\"א 6688101",
  "status": "valid",
  "isValid": true,
  "needsReview": false,
  "cityCode": 5000,
  "cityHe": "תל אביב - יפו",
  "cityEn": "TEL AVIV - YAFO",
  "subdistrict": "תל אביב",
  "cityMatchType": "exact",
  "streetCode": 1103,
  "streetHe": "שדרות רוטשילד",
  "streetMatchType": "without_street_type",
  "streetSimilarity": 1,
  "houseNumber": "1",
  "postalCode": "6688101",
  "normalizedAddressHe": "שדרות רוטשילד 1, תל אביב - יפו",
  "suggestions": [],
  "issues": ["Postal code copied as written; it is not validated (not part of the source lists)"],
  "sourceUpdatedAt": "2026-11-01T01:30:34.891058"
}

Statuses: valid, valid_street_no_house_number, valid_locality_without_streets (the official list has no named streets for that locality), street_not_found, ambiguous_street, street_missing, city_not_found, ambiguous_city, empty_input.

A SUMMARY record in the key-value store holds counts by status, the source resources with their last-modified dates, and the attribution line.

How to use

1. Paste addresses into addresses (one per line) or upload JSON records with your own ids.
2. Start the run. Loading the official lists takes a few seconds; the matching itself is fast.
3. Open the Problems view of the dataset to see only rows that need a human look, or export everything as CSV/Excel/JSON.

Measured locally: 5,000 addresses took about 9 seconds end to end, including about 6–7 seconds to download the lists. Run time on the Apify platform depends on the memory you choose.

Accuracy (self-test, with caveats)

bench/benchmark.mjs builds addresses from 2,000 random rows of the official list and checks whether the Actor returns the same locality and street codes (seed 42, run on 2 Nov 2026):

• Official spelling, street number, city: 100% same codes.
• With רח' prefix and no comma: 99.4% same codes; 11 marked ambiguous, 1 wrong (flagged for review).
• City first: 100% same codes.
• One letter deleted from the street name: 96.9% same codes; 41 marked ambiguous; 22 wrong, of which 3 were not flagged, because the typo turned the name into another real street name in the same locality.

These numbers come from addresses generated from the official list itself. They do not measure real customer data, which is messier. Always review rows with needsReview: true.

Pricing (suggestion for the publisher, not final)

Pay per event: one address-validated event per address checked, suggested at about $1.50 per 1,000 addresses. With onlyProblems every checked address is still charged. The run stops cleanly when your maximum cost per run is reached.

Limits

• Street names must be in Hebrew. English is supported for locality names only.
• House numbers are not checked against a building list (the source has none). A valid status means the locality and the street exist officially.
• Postal codes are not validated, only copied.
• Some small localities have no named streets in the official list; these return valid_locality_without_streets.
• Local or informal street names, new streets not yet in the published list, and P.O. boxes are not recognised.
• Arabic-script addresses are not supported.

Data source, license and legal note

• Data: "רשימת ישובים בישראל" and "רשימת רחובות ישראל" on data.gov.il, published by the Population and Immigration Authority. Source: data.gov.il, Population and Immigration Authority (רשות האוכלוסין וההגירה).
• License: the data.gov.il usage license allows commercial use, copying and derived works, and requires naming the source. The localities dataset is marked "Other (Open)"; the streets dataset has no separate license, so the site license applies.
• The lists hold locality and street names and codes only. No personal data is downloaded. Your input addresses are processed only inside your own run and stored only in your run's dataset.
• This Actor is not affiliated with or endorsed by the Government of Israel or the Population and Immigration Authority.
• Results are an automated check against public lists. They are not a guarantee that mail or couriers can deliver to an address, and they are not legal advice.

Support

Open an issue on the Actor's Issues tab with an example address (remove personal details) and the output you expected.