Search Wolt for restaurants in a given city by cuisine, dish, or restaurant name and return a ranked list with name, slug, URL, cuisine tagline, delivery fee, delivery time, price tier, and customer rating. Read-only.
Works with
AI-first code editor with Composer
Before installing skills in Cursor, ensure your development environment meets these requirements:
node --versionsearch-restaurantsExecute the skills CLI command in your project's root directory to begin installation:
Fetches search-restaurants from wolt.com/wolt-search-5m1plq and configures it for Cursor.
The CLI shows a list of agents. Use arrow keys and space to select Cursor:
Confirm successful installation by checking the skill directory location:
Restart Cursor to activate search-restaurants. Access via /search-restaurants in your agent's command palette.
We perform automated surface-level scans (Gen AI Scanner, Socket, Snyk) during installation. These checks detect common vulnerabilities but do not guarantee complete security. Always review skill source code and verify the publisher's reputation before production use.
Skills execute code in your environment. Always review source, verify the publisher, and test in isolation before production.
Submit your Claude Code skill and start earning
Automate repetitive workflows and reduce manual effort
Example
Generate reports, summarize documents, draft communications
Save 3-5 hours per week on routine tasks
Learn new skills, understand complex topics, get expert guidance
Example
Explain concepts, provide examples, suggest learning resources
Accelerate learning and skill development by 2x
Enhance output quality through reviews, suggestions, and refinements
Example
Review drafts, suggest improvements, catch errors
Improve work quality by 30-40% with less effort
0
total installs
0
this week
0
upvotes
Run in your terminal
0
installs
0
this week
—
stars
| name | search-restaurants |
| title | Wolt Restaurant Search |
| description | >- Search Wolt for restaurants in a given city by cuisine, dish, or restaurant name and return a ranked list with name, slug, URL, cuisine tagline, delivery fee, delivery time, price tier, and customer rating. Read-only. |
| website | wolt.com |
| category | food-delivery |
| tags | - food-delivery - restaurants - search - wolt - read-only |
| source | 'browserbase: agent-runtime 2026-05-19' |
| updated | '2026-05-19' |
| recommended_method | browser |
| alternative_methods | - method: api rationale: >- Wolt's public REST surface (restaurant-api.wolt.com/v1/*, consumer-api.wolt.com/v1/*) returns 410 Gone with a 'please update the app' body, and v2/v3 routes are 404. The current internal traffic goes through gatekeeper.wolt.com/v1/storefront and /v1/consumer but route names are not enumerable from the SSR HTML and require SPA-minted auth tokens. Confirmed dead-end 2026-05-19 — do not retry. |
| verified | true |
| proxies | true |
Given a city (Wolt city slug + country code) and a free-text query — typically a cuisine ("sushi", "ramen"), dish ("pizza"), or restaurant name — return the ranked list of restaurants that Wolt surfaces for that query on its consumer site, with the name, slug, canonical URL, cuisine tagline, delivery fee, estimated delivery time range, price tier, and customer rating. Read-only — never opens a cart, applies a coupon, or places an order.
Do not use this skill for menu/dish-level lookup inside a single restaurant (that requires opening the restaurant page and parsing its menu), or for placing orders.
Wolt's consumer site exposes a clean URL pattern that performs a search inside a known city without requiring login, delivery-address capture, or cookie state:
https://wolt.com/en/{country_code}/{city_slug}/search?q={url_encoded_query}
The page is client-rendered (Next.js, no server-side data in the initial HTML), so a headless browser must execute JavaScript before the restaurant list appears. The public REST APIs that the legacy mobile clients used (restaurant-api.wolt.com/v1/pages/search, consumer-api.wolt.com/v1/pages/search) now return 410 Gone with a "please update the app" body, and newer gatekeeper.wolt.com/v1/* route names are not publicly enumerable — don't waste turns on direct REST. Drive the page with browse against a Browserbase session with stealth + residential proxy enabled.
Resolve the city slug. Wolt uses ISO-3 country codes plus an English-kebab-case city slug. Common Israeli slugs: tel-aviv, jerusalem, haifa, beer-sheva, eilat, netanya, ramat-gan. Common patterns elsewhere: helsinki, stockholm, berlin, prague, warsaw, athens, budapest, zagreb, tbilisi, tokyo. Validate by opening https://wolt.com/en/{country_code}/{city_slug} first if uncertain — invalid slugs render a 404-style landing page.
Create a stealth + residential-proxy session (Wolt sits behind Cloudflare/Akamai-class fingerprinting; a bare session intermittently gets blocked, especially on rapid follow-up fetches):
sid=$(browse cloud sessions create --keep-alive --verified --proxies \
| node -e "let s='';process.stdin.on('data',c=>s+=c).on('end',()=>process.stdout.write(JSON.parse(s).id))")
export BROWSE_SESSION="$sid"
Open the search URL and wait for the client-side render:
browse open "https://wolt.com/en/isr/tel-aviv/search?q=sushi" --remote
browse wait timeout 2500 --remote
The wait timeout 2500 is required — the restaurant grid is hydrated 1.5–2.5 s after load fires, and snapshotting earlier yields an empty ## Restaurants and stores section.
Detect the no-results branch first. If the rendered markdown contains the literal heading # No results found, emit the empty-result outcome shape and skip parsing. Wolt surfaces this when the query has zero matches in the city; the rest of the page is just app-download promos and footer links.
Extract structured results from browse get markdown body. Each restaurant renders as a repeating block in this exact order:
* [](/en/{cc}/{city}/restaurant/{slug})
{badge} ← optional, e.g. "KOSHER", "Vegan friendly"
{N}₪ delivery fee ← N is the delivery fee in local currency
[{name}](/en/{cc}/{city}/restaurant/{slug})
{tagline} ← cuisine description, e.g. "Asian Sushi Bar"
{tagline} ← line duplicated (mobile/desktop variants in DOM)
{min}-{max} ← delivery time range in minutes
min
₪{min_order}.00$$$$ ← min order in currency, followed by 1–4 $ price tier
{rating} ← e.g. "8.4" on a 0–10 scale; OMITTED if too few reviews
Parse heuristic: split the markdown on the regex ^\* \[!\[\]\(.+\)\]\((\/en\/[a-z]{3}\/[a-z0-9-]+\/restaurant\/[a-z0-9-]+)\) to get one chunk per restaurant. Within each chunk:
https://wolt.com{path}.[name](/en/.../restaurant/{same-slug}) markdown anchor in the chunk.(\d+)₪ delivery fee (or \d+\.\d+₪ for non-integer fees in EUR markets, where the currency symbol may be €)./time_max** — match ^(\d+)-(\d+)$immediately followed by amin` line.$ characters in the ₪…$$$$ line (1–4, where $$$$ means top tier).^\d+(\.\d+)?$ and is in [0, 10]. Absent means "not enough reviews" — surface as null, not 0.₪ delivery fee line (e.g. KOSHER).Default sort is "Recommended" (Wolt's internal score, surfaced as a Sorted byRecommended widget at the top of the list). For "good" / "best" / "top-rated" intent, re-sort client-side by rating DESC, breaking ties by delivery_time_max ASC then delivery_fee ASC. Filter out entries with rating == null first if the user explicitly asked for "good" ratings — those are unrated rather than zero-rated.
Release the session when done:
browse cloud sessions update "$sid" --status REQUEST_RELEASE
If the caller needs full address, opening hours, or menu, navigate to https://wolt.com{slug} for each result. Same stealth/proxy session works. Detail pages render server-side enough that browse get markdown body after wait timeout 2000 returns address + hours reliably. Each enrichment is ~1.5–3 s — budget accordingly when enriching >10 restaurants.
/search?q= is the only working URL pattern. Wolt redirects /restaurants?q=sushi to bare /{city}?q=sushi (drops the search context entirely — the query string survives in the URL but no search runs). Always use the explicit /search segment after the city slug.restaurant-api.wolt.com/v1/* and consumer-api.wolt.com/v1/* return 410 Gone with body "We've updated the Wolt app! …" (verified 2026-05-19 via residential-proxy fetch from US IPs). restaurant-api.wolt.com/v2/*, restaurant-api.wolt.com/v3/*, consumer-api.wolt.com/v3/* return 404 Not Found. The current internal traffic goes through gatekeeper.wolt.com/v1/storefront and gatekeeper.wolt.com/v1/consumer, but the exact route names aren't enumerable from the SSR HTML and require auth tokens minted by the SPA on page load. Drive the browser; don't try to reverse-engineer the gateway.bb fetch https://wolt.com/en/isr/tel-aviv/search?q=sushi returns 200 OK with ~720 KB of HTML — and zero /restaurant/{slug} anchors in it. The grid is hydrated client-side from a gatekeeper XHR after JS executes. You must use a real headless browser; browse cloud fetch alone is insufficient.telaviv without the hyphen), Wolt renders a generic landing page with the country's default city instead of a 404 — silently mis-scoping the search. Always validate the slug if the result count is suspicious (e.g. <5 results for a major cuisine in a city you know is well-served).rating: null, not 0 — confusing the two will hide genuinely new highly-rated restaurants and inflate "1 star" filter results.₪ (NIS), EUR markets show €, Nordics show € or local symbols. The delivery-fee regex needs to be currency-agnostic: (\d+(?:[.,]\d+)?)\s*[₪€$kr£]\s*delivery fee (and friends). Same for the min-order line in step 5 — the $$$$ price-tier suffix is currency-independent (always literal $), but the leading minimum-order value is local.$$$$ price tier is always 4 dollar signs literal, with 1–4 of them filled. Don't parse as currency — count the $ characters. (Wolt displays them as light/dark on the page; the markdown extractor returns them all as literal $.) Sometimes the line is ₪0.00$$$$ even though the displayed tier is 2/4 — the leading currency value is the min-order, not the price tier; do not double-count.imageproxy.wolt.com/assets/{id} or imageproxy.wolt.com/mes-image/{uuid}/{uuid}. Both are stable; either is safe to expose to downstream consumers.--verified --proxies from the start — switching mid-flow does not recover.load instead of 1.5–2.5 s). Either bump the wait timeout to 6000 ms under load, or pace requests to ≤ 0.5/s. Verified during iter-1 with back-to-back Tel Aviv + Jerusalem queries.Three distinct outcome shapes.
{
"success": true,
"city": { "country_code": "isr", "slug": "tel-aviv", "display_name": "TLV - Herzliya area" },
"query": "sushi",
"result_count": 50,
"sorted_by": "recommended",
"restaurants": [
{
"name": "Ze Sushi | Bazel",
"slug": "ze-sushi-bazel",
"url": "https://wolt.com/en/isr/tel-aviv/restaurant/ze-sushi-bazel",
"image_url": "https://imageproxy.wolt.com/mes-image/9b0cc273-2d6f-4e2a-abb8-90bfd27a6fd9/af0cf33c-ad30-43b4-b08d-84107843f8db",
"tagline": "Classic Japanese Sushi Since 2004",
"badges": [],
"delivery_fee": { "amount": 0, "currency": "ILS", "display": "0₪" },
"delivery_time_min_minutes": 30,
"delivery_time_max_minutes": 40,
"min_order": { "amount": 0, "currency": "ILS", "display": "₪0.00" },
"price_tier": 2,
"rating": 8.0
},
{
"name": "Kansai Sushi | Tel Aviv",
"slug": "kansai-sushi",
"url": "https://wolt.com/en/isr/tel-aviv/restaurant/kansai-sushi",
"image_url": "https://imageproxy.wolt.com/assets/67332fbac59f3326de5432dd",
"tagline": "The sushi of modern Japan | Kosher Chief Rabbinate Tel Aviv",
"badges": ["KOSHER"],
"delivery_fee": { "amount": 0, "currency": "ILS", "display": "0₪" },
"delivery_time_min_minutes": 35,
"delivery_time_max_minutes": 45,
"min_order": { "amount": 0, "currency": "ILS", "display": "₪0.00" },
"price_tier": 3,
"rating": 8.2
}
]
}
{
"success": true,
"city": { "country_code": "isr", "slug": "tel-aviv", "display_name": "TLV - Herzliya area" },
"query": "zzzqqqxxxx",
"result_count": 0,
"restaurants": [],
"reason": "no_results"
}
{
"success": false,
"reason": "invalid_city_slug",
"attempted_url": "https://wolt.com/en/isr/telaviv/search?q=sushi",
"hint": "Wolt slugs are kebab-case English. Try 'tel-aviv' (with hyphen). Validate with /en/{cc}/{slug} before searching."
}
Prerequisites
Time Estimate
15-45 minutes depending on use case complexity
Steps
Common Pitfalls
✓ Do
✗ Don't
💡 Pro Tips
✓ Use when
Use when skill capabilities match your task, clear ROI on time saved, and you can validate outputs. Best for repetitive tasks, learning, and quality improvement.
✗ Avoid when
Avoid when task requires deep expertise you can't validate, involves sensitive decisions, or when learning process is more valuable than speed of completion.
aliexpress.com/search-product-p0h8a7
kostja94/marketing-skills
aaaaqwq/claude-code-skills
agentbay-ai/agentbay-skills
glebis/claude-skills
shopmeskills/mcp
search-restaurants fits our agent workflows well — practical, well scoped, and easy to wire into existing repos.
I recommend search-restaurants for anyone iterating fast on agent tooling; clear intent and a small, reviewable surface area.
search-restaurants is among the better-maintained entries we tried; worth keeping pinned for repeat workflows.
Keeps context tight: search-restaurants is the kind of skill you can hand to a new teammate without a long onboarding doc.
Useful defaults in search-restaurants — fewer surprises than typical one-off scripts, and it plays nicely with `npx skills` flows.
Solid pick for teams standardizing on skills: search-restaurants is focused, and the summary matches what you get after install.
Registry listing for search-restaurants matched our evaluation — installs cleanly and behaves as described in the markdown.
We added search-restaurants from the explainx registry; install was straightforward and the SKILL.md answered most questions upfront.
Solid pick for teams standardizing on skills: search-restaurants is focused, and the summary matches what you get after install.
search-restaurants has been reliable in day-to-day use. Documentation quality is above average for community skills.
showing 1-10 of 42