OpenRice Scraper | Hong Kong Restaurant Data, Ratings, Districts & Cuisines (No API)

Scrape restaurant search listings from OpenRice Hong Kong, including restaurant name, cuisine, district, rating, review count, address, phone number, image URLs and more.

• Public JSON API — no login, no token, no browser required
• The only stable option: the sole competitor on Apify Store has been stuck under UNDER_MAINTENANCE for a long time, with 30/38 runs timing out
• Bulk keyword search: run multiple keywords in one job, with optional district / cuisine ID filters
• Structured output: snake_case fields, missing values filled with null, every row stamped with an ISO 8601 timestamp

---

Output fields

Field	Type	Description	Version
poi_id	string	OpenRice restaurant unique ID	v1.0
name	string | null	Restaurant name (Chinese)	v1.0
cuisine	string[] | null	Cuisine array (e.g. ["Japanese", "Sushi/Sashimi"])	v1.0
district	string | null	District name (e.g. "Causeway Bay")	v1.0
price_range	string | null	Price range: $ / $$ / $$$ / $$$$	v1.0
rating	number | null	OpenRice overall rating (e.g. 3.87)	v1.0
review_count	integer | null	Number of reviews	v1.0
bookmark_count	integer | null	Number of users who bookmarked the restaurant	v1.0
address	string | null	Restaurant address (Chinese)	v1.0
phone	string | null	Phone number (first one)	v1.0
opening_hours	string | null	Opening hours as a JSON string (v1.1 will switch to a human-readable format)	v1.0
image_urls	string[]	Array of restaurant image URLs ([] if none)	v1.0
restaurant_url	string	Full URL of the OpenRice restaurant page	v1.0
source_mode	string	Data source mode (currently "search")	v1.0
source_query	string	The search keyword that produced this row	v1.0
scraped_at	string	Scrape timestamp (ISO 8601, UTC)	v1.0

---

Modes

mode	Input	Output	Billing event
search (v1.0)	keywords[] + optional district / cuisine filters	Search listing results	search-listing ($0.002 per row)
detail (v1.1, in development)	restaurantIds[] or restaurantUrls[]	Full restaurant detail	product-detail

---

Input parameters

Parameter	Type	Required	Default	Description
keywords	string[]	✓	—	Search keywords (restaurant name / cuisine), up to 50
bulkInput	string	—	—	One keyword per line, automatically merged into keywords
district	string	—	—	District ID (e.g. 1019 = Causeway Bay, see below)
cuisine	string	—	—	Cuisine ID (e.g. 2009 = Japanese, see below)
maxItems	integer	—	50	Max results per keyword (1-500)
proxyConfiguration	object	—	disabled	Apify Proxy or a custom proxy

Common district IDs

districtId	District
1999	Hong Kong Island (all)
1019	Causeway Bay
1001	Sheung Wan / Central
2999	Kowloon (all)
2010	Mong Kok
2001	Tsim Sha Tsui
3999	New Territories (all)

For the full list, call GET https://www.openrice.com/api/v2/metadata/region/all?uiLang=zh&uiCity=hongkong.

Common cuisine IDs

cuisineId	Cuisine
1004	Cantonese / HK-style
2009	Japanese
4000	Western
1007	Chinese
2001	Korean

---

Input examples

1. Search for sushi (default settings)

{
    "keywords": ["sushi"],
    "maxItems": 20
}

2. BBQ filtered to Central/Sheung Wan (districtId=1001)

{
    "keywords": ["bbq"],
    "district": "1001",
    "maxItems": 20
}

3. Multiple keywords in one run

{
    "keywords": ["japanese food", "korean bbq"],
    "maxItems": 50
}

4. Japanese restaurants in Mong Kok (district + cuisine filter combined)

{
    "keywords": ["sushi"],
    "district": "2010",
    "cuisine": "2009",
    "maxItems": 30
}

5. Bulk-paste multiple keywords

{
    "bulkInput": "sushi\nbbq\nhot pot\ncha chaan teng\ndim sum",
    "maxItems": 30
}

---

Notes

• Reliability is the selling point: the only competitor has been stuck UNDER_MAINTENANCE for a long time (30/38 runs timing out); this Actor uses a pure JSON API, no browser, and exponential-backoff retries — finishing the job is the whole point
• Billing timing: Actor.charge() is only called after data has been successfully pushed to the dataset; failures mid-run are never billed
• Consecutive empty-page detection: the run auto-stops if 5 consecutive pages return no new data, avoiding wasted compute
• detail mode (v1.1): OpenRice's POI detail API endpoint is not currently public — all probe requests return 404; will be added once found or reverse-engineered

---

FAQ

Q: Why are some fields null? A: Some OpenRice restaurant listings have incomplete data (e.g. no phone number, no address filled in). This Actor always fills missing values with null instead of dropping the field, so every row keeps a consistent structure.

Q: What format is opening_hours in? A: In v1.0 it's a JSON string containing per-day fields like dayOfWeek (1=Sunday, 2=Monday...7=Saturday), period1Start, period1End, is24hr, isClose, etc. v1.1 will switch to a human-readable format.

Q: Can I search a specific district only? A: Yes — fill in district with a districtId (e.g. 1019 = Causeway Bay). The full districtId list is available from the metadata API.

Q: Is proxy supported? A: Yes, both Apify Proxy and a custom proxy are supported; disabled by default since direct connections already scrape reliably.

Q: Does image_urls include all photos? A: It includes the restaurant's main photo (doorPhoto) plus the latest images returned in the search listing (usually 3-5). The full photo album requires detail mode (v1.1).

Q: Can I scrape review content? A: A reviews mode is planned for v1.1 and is not yet implemented.

---

Deploy

npm install
npm run build
apify push

Local testing:

$apify run --input='{"keywords":["sushi"],"maxItems":5}'

---

Disclaimer

This tool is intended only for lawful research and analysis of publicly accessible data. Users must comply with OpenRice's Terms of Service, avoid mass scraping, avoid use for commercial competition, and avoid storing personal data. The author bears no responsibility for how users choose to use this tool.

---

OpenRice scraper, Hong Kong restaurants, HK dining data, restaurant ratings, cuisine search, OpenRice API, Hong Kong food data