# Watch Arbitrage — Patek/Rolex/AP Cross-Platform Alerts (`kazkn/watch-arbitrage-mcp`) Actor Track Patek Philippe, Rolex & Audemars Piguet price spreads across 6 marketplaces (Chrono24, WatchBox, Bobs, Watchfinder UK, European Watch, Watches of Switzerland). Real-time Telegram alerts when a listing is mispriced cross-platform. Filter by condition + box/papers. Built for pro watch dealers. - **URL**: https://apify.com/kazkn/watch-arbitrage-mcp.md - **Developed by:** [KazKN](https://apify.com/kazkn) (community) - **Categories:** E-commerce, AI, MCP servers - **Stats:** 1 total users, 0 monthly users, 0.0% runs succeeded, NaN bookmarks - **User rating**: No ratings yet ## Pricing from $1,500.00 / 1,000 spread alert triggereds This Actor is paid per event and usage. You are charged both the fixed price for specific events and for Apify platform usage. Since this Actor supports Apify Store discounts, the price gets lower the higher subscription plan you have. Learn more: https://docs.apify.com/platform/actors/running/actors-in-store#pay-per-event ## What's an Apify Actor? Actors are a software tools running on the Apify platform, for all kinds of web data extraction and automation use cases. In Batch mode, an Actor accepts a well-defined JSON input, performs an action which can take anything from a few seconds to a few hours, and optionally produces a well-defined JSON output, datasets with results, or files in key-value store. In Standby mode, an Actor provides a web server which can be used as a website, API, or an MCP server. Actors are written with capital "A". ## How to integrate an Actor? If asked about integration, you help developers integrate Actors into their projects. You adapt to their stack and deliver integrations that are safe, well-documented, and production-ready. The best way to integrate Actors is as follows. In JavaScript/TypeScript projects, use official [JavaScript/TypeScript client](https://docs.apify.com/api/client/js.md): ```bash npm install apify-client ``` In Python projects, use official [Python client library](https://docs.apify.com/api/client/python.md): ```bash pip install apify-client ``` In shell scripts, use [Apify CLI](https://docs.apify.com/cli/docs.md): ````bash # MacOS / Linux curl -fsSL https://apify.com/install-cli.sh | bash # Windows irm https://apify.com/install-cli.ps1 | iex ```bash In AI frameworks, you might use the [Apify MCP server](https://docs.apify.com/platform/integrations/mcp.md). If your project is in a different language, use the [REST API](https://docs.apify.com/api/v2.md). For usage examples, see the [API](#api) section below. For more details, see Apify documentation as [Markdown index](https://docs.apify.com/llms.txt) and [Markdown full-text](https://docs.apify.com/llms-full.txt). # README ## Watch Arbitrage Tracker (Patek/Rolex/AP) ![Watch Arbitrage Tracker — Spot mispriced Patek, Rolex & AP across 6 marketplaces](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/01-hero.png) > Cross-platform price arbitrage tool for **professional watch dealers** trading **Patek Philippe**, **Rolex**, and **Audemars Piguet (AP)** across **Chrono24**, **WatchBox** (now part of **The 1916 Company** — the 2024 merger of WatchBox + **Govberg Jewelers** + **Hyde Park Jewelers**), **Bobs Watches**, **Watchfinder UK**, **European Watch Co (Boston)**, and **Watches of Switzerland (Mayors / Mappin & Webb / Goldsmiths group)**. **Real-time Telegram alerts** when a listing is mispriced cross-platform. ### What does Watch Arbitrage Tracker do? This Actor monitors a list of watch references you care about (e.g. `5711/1A`, `116500LN`, `15500ST`) across **6 major dealer marketplaces simultaneously**, computes the **cross-platform median price**, and fires a **Telegram alert** the moment any listing drops more than `X%` (default 5%) below the cross-platform median. It runs on **Apify proxy + Camoufox** (a stealthy Firefox fork) so Cloudflare-protected sites like Chrono24 and Bobs Watches are scraped reliably. Schedule it once an hour and you have a real-time arbitrage radar — packaged as a transparent **Pay-Per-Event** Actor. #### Platform coverage (verified 2026-05-06, build 0.1.19) | Source | Status | Notes | |---|---|---| | **Chrono24** | ✅ Confirmed delivering listings | Largest global marketplace; widest variance — most arbitrage signals come from here | | **WatchBox / The 1916 Company** | ✅ Confirmed delivering listings | Premium US dealer (ex-Govberg + ex-Hyde Park merger 2024). Domain: `the1916company.com` (legacy `watchbox.com` redirects) | | **Bobs Watches** | ✅ Confirmed delivering listings | Rolex specialist, Newport Beach CA. Cloudflare-protected — Camoufox handles | | **Watchfinder UK** | ✅ Confirmed delivering listings | UK leader (Richemont Group). Cross-platform median signal validated | | European Watch Co (Boston) | ⚠️ Discovery in progress (v0.2) | 69+ Patek inventory verified but ref-grid filter calibration ongoing | | Watches of Switzerland | ⚠️ Discovery in progress (v0.2) | Patek pricing hidden ("price on request"); Rolex/AP works when in stock | **4 platforms confirmed delivering cross-platform median signals** as of build 0.1.19. The 2 remaining sources are wired and run on every scheduled batch — when they return inventory matching your refs, listings flow into the dataset automatically. ### Why use Watch Arbitrage Tracker? - **Stop tab-switching**: 6 sources (US + UK + EU coverage) monitored on every run, no manual refresh. - **Find spreads other dealers miss**: a Daytona listed cheap on WatchBox while three other dealers list it $5K higher on Chrono24 = instant flip. We surface it before it sells. - **TRUE cross-platform median**: median is computed across all sources that returned listings for your sub-ref — not single-platform anchoring like Watchcharts or Bezel Club. - **EU↔US arbitrage built-in**: simplified import/VAT estimate (+20%) on EU listings so US dealers know the true landed cost. - **Telegram-native**: dealers live on Telegram. Alerts land in your group chat with brand+sub-ref, asking price, median, spread % below median, and a one-tap link to the listing. - **Pay only when you win**: PPE pricing — you pay $0.05 per ref/day to monitor + $0.50 *only* when an actual spread is detected and a real Telegram alert is dispatched. ### How to use Watch Arbitrage Tracker #### Step 1 — Quick Start (just 3 fields) Pick the watches to monitor + alert sensitivity + alert channel. Smart defaults handle the rest. ![Quick Start input section — references, spread sensitivity, alert channel](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/02-input-quick-start.png) #### Step 2 — Telegram bot setup (one-time, 2 minutes) Open Telegram, search for **@BotFather** (verified by Telegram with a blue check): ![@BotFather verified bot profile in Telegram](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/06-botfather-profile.png) Tap *Message* → type `/newbot` → choose any name (e.g. "Watch Arbitrage"): ![BotFather /newbot flow — name your new bot](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/07-botfather-newbot.png) BotFather will reply with your **bot token** (a long string like `8602184201:AAEa...`). Copy it. Then paste it into the Telegram setup section of the input form, along with your numeric chat ID (get it via `@userinfobot`): ![Telegram setup — bot token + chat ID input fields](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/03-input-telegram-setup.png) #### Step 3 — (Optional) Filters Refine which listings count toward arbitrage detection — by condition (new / pre-owned / vintage) and box-and-papers status (full-set / box-only / etc.): ![Filter listings by condition and box-and-papers status](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/04-input-filters.png) #### Step 4 — (Optional) Advanced Defaults work great for most dealers. Tweak only if you have a specific need (e.g. exclude UK/EU sources): ![Advanced options — marketplaces and listing cap](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/05-input-advanced.png) #### Step 5 — Click Start (or schedule it) The Actor runs in ~5 minutes. Telegram pings on real spreads with asking price + cross-platform median + a one-tap listing link. > **Tip — recommended ref portfolio size:** the more refs you monitor, the more cross-platform signals surface. 3-5 refs proves the concept; 15-30 refs catches real flippable spreads multiple times per week. Volume dealers run 50+ refs. ### Input The Apify Console form is grouped in sections (Quick Start, Telegram setup, Filters, Advanced). Smart defaults work for most dealers. | Field | Type | Description | |---|---|---| | `references` | string[] | **Precise reference numbers recommended** (e.g. `5711/1A-010`, `116500LN`, `15500ST.OO.1220ST.04`). Model names like `Nautilus`, `Daytona`, `Royal Oak` are accepted but produce less precise spread alerts (sub-models conflated). | | `spread_sensitivity` | enum string | `"3"` to `"20"` — % below cross-platform median to trigger alert. Default `"5"` (recommended balance). | | `alert_channel` | enum | `telegram` / `dataset_only`. Default `telegram`. *(Email digest coming v0.2 — see "What's new" for details on why it's deferred.)* | | `alert_telegram_bot_token` | string (secret) | Bot token from `@BotFather`. Required for Telegram. | | `alert_telegram_chat_id` | string | Numeric chat ID via `@userinfobot`. | | `filter_conditions` | enum[] | Only include listings with these conditions: `new`, `like-new`, `very-good`, `good`, `fair`, `vintage`, `pre-owned`, `unknown`. | | `strict_condition_matching` | bool | When `true`: exclude listings where condition couldn't be reliably scraped. Pair with Bobs + Watchfinder + European Watch for clean filtering. | | `filter_box_papers` | enum[] | `full-set`, `box-and-papers`, `papers-only`, `box-only`, `watch-only`, `unknown`. | | `platforms` | enum[] | Default: all 6 active. Hodinkee deprecated (acquired by WoS Group, inventory winding down) but kept in enum for backward-compat. | | `max_listings_per_ref_per_platform` | int | Cap to control compute cost (default 25). | Example input: ```json { "references": ["5711/1A-010", "116500LN", "124060", "15500ST.OO.1220ST.04"], "platforms": ["chrono24", "watchbox", "bobs", "watchfinder", "europeanwatch", "watchesofswitzerland"], "spread_sensitivity": "5", "alert_channel": "telegram", "alert_telegram_chat_id": "123456789", "filter_conditions": ["new", "like-new", "very-good", "pre-owned"], "strict_condition_matching": false } ```` ### Output Three artifacts per run: 1. **Dataset** — every individual listing scraped (CSV / JSON / Excel export). 2. **`MARKET_SNAPSHOT`** (KV store) — per-reference median, min, max, count across all 6 platforms. 3. **`ARBITRAGE_OPPORTUNITIES`** (KV store) — only the listings that crossed the spread threshold. #### Dataset table view (Apify Console "Output" tab) Every individual listing scraped, with reference, brand, marketplace, price (USD-normalized), original currency, year, dealer, location, and direct listing URL. Sortable, filterable, exportable as JSON / CSV / Excel: ![Dataset table view — every listing scraped across all platforms](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/09-output-dataset-table.png) #### `ARBITRAGE_OPPORTUNITIES` (key-value store) — only the spreads that fired Real arbitrage opportunities from a build 0.1.18 cloud run, formatted as JSON. Includes the full `listing` object plus the cross-platform `median_usd` + `spread_usd` + `spread_pct`: ![ARBITRAGE\_OPPORTUNITIES key-value store — real spreads detected](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/10-output-arbitrage-opportunities.png) #### `MARKET_SNAPSHOT` (key-value store) — per-sub-ref aggregate stats Median / min / max / count + the list of platforms that contributed listings, for every sub-ref detected on the run: ![MARKET\_SNAPSHOT key-value store — per-sub-ref median/min/max](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/11-output-market-snapshot.png) #### Apify Console run summary Every run lands in the Console with a clean status header (Succeeded / Failed), runtime, compute units, total cost, and links to the dataset + KV store + live view URL: ![Apify Console run summary — Succeeded, 30/0, $0.167, 7m 58s](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/12-console-run-summary.png) Sample dataset item (Patek 5711/1A-010 listed on Chrono24): ```json { "ref": "5711/1A-010", "brand": "patek-philippe", "platform": "chrono24", "title": "Patek Philippe Nautilus 5711/1A-010 - Calibre 324 S C", "price_usd": 110982, "price_orig": 102450, "currency": "EUR", "listing_url": "https://www.chrono24.com/patekphilippe/nautilus--id45152307.htm", "dealer": "European Watch Co.", "condition": "Very good", "condition_normalized": "very-good", "box_papers": "full-set", "year": 2018, "location": "DE", "scraped_at": "2026-05-05T15:55:00.000Z" } ``` Sample arbitrage opportunity (real result from build 0.1.18 cloud run, 2026-05-06): ```json { "ref": "124060", "brand": "rolex", "median_usd": 13988, "spread_usd": 3938, "spread_pct": 28.2, "listing": { "ref": "124060", "platform": "watchbox", "title": "Rolex Submariner No Date 124060", "price_usd": 10050, "dealer": "1916 Company (ex-WatchBox)", "listing_url": "https://www.the1916company.com/.../124060-..." }, "detected_at": "2026-05-06T00:46:42.000Z" } ``` This was a real cross-platform spread: WatchBox listed a Submariner 124060 at $10,050 while the median across Chrono24 + Bobs + WatchBox was $13,988 — a 28.2% spread, ~$3,900 below market in plain dollars. #### Telegram alert format (real screenshots from a live run) This is what actually arrived in our test Telegram chat from the build 0.1.19 run on 2026-05-06 — three real arbitrage opportunities on a Rolex Submariner Kermit (`126610LV`): ![Real Telegram alerts — 3 spreads detected on Rolex 126610LV](https://raw.githubusercontent.com/DataKazKN/watch-arbitrage-mcp/main/actor/assets/screenshots/08-telegram-alerts-real.png) Each alert includes: - **Brand + sub-reference** (`Rolex — 126610LV`) - **Asking price** + the platform name (`$11,900 on Chrono24` — clean platform labels, not raw enums) - **Cross-platform median** computed across every source that returned listings - **Spread** in both dollars and % below median - **Year** + dealer fields when scraped reliably - **Asking price disclaimer** — listed price excl. shipping, taxes, customs (varies by buyer location) - **Direct link** to the listing on the source platform > The asking-price disclaimer is important: dealer-listed price ≠ delivered price. Add ~10-20% for shipping + import duty + sales tax depending on your buyer's geography. The alert reflects the *asking price the dealer typed on their listing page* — that's the price you negotiate from. Note: arbitrage stats are computed per **sub-reference** (e.g. `5711/1A-010` vs `7011/1G` are tracked as separate sub-models even when both come from a `Nautilus` search). This eliminates false positives where a ladies' Nautilus would flag as a "spread" against the men's median. ### Data table — fields per listing | Field | Description | |---|---| | `ref` | Reference number you queried | | `brand` | Auto-detected: `patek-philippe` / `rolex` / `audemars-piguet` | | `platform` | `chrono24` / `watchbox` / `bobs` / `watchfinder` / `europeanwatch` / `watchesofswitzerland` | | `title` | Listing title from the source page | | `price_usd` | Normalized to USD via static FX (EUR/GBP/CHF/etc.) | | `price_orig` | Raw price as listed | | `currency` | Original currency | | `condition` | Raw condition string from source page (e.g. `"Very good"`, `"Pre-owned"`) | | `condition_normalized` | Normalized to `new` / `like-new` / `very-good` / `good` / `fair` / `vintage` / `pre-owned` / `unknown` | | `box_papers` | Detected box & papers status: `full-set` / `box-and-papers` / `papers-only` / `box-only` / `watch-only` / `unknown` | | `dealer` | Dealer / seller name | | `year` | Manufacture year if available | | `location` | Country code if available | | `listing_url` | Direct link to the listing | | `scraped_at` | ISO timestamp | ### Pricing Pay-Per-Event — you only pay for what the Actor actually does. See current pricing on the **Apify Store page** (live rates and plan-based discounts shown there). Four chargeable events: | Event | When | |---|---| | `actor-start` | Once per run | | `reference-monitored` | Per reference scanned across all 6 platforms | | `apify-default-dataset-item` | Per listing scraped | | `spread-alert-triggered` | **Primary event** — only when a real arbitrage opportunity is dispatched (Telegram or email) | **Discounts apply to higher Apify plans** (Bronze / Silver / Gold) — visible on the Store page pricing tab. **Spending limit respected**: the Actor reads `ACTOR_MAX_TOTAL_CHARGE_USD` and stops gracefully when your run budget is reached, so a runaway alert spike never blows past the cap you set in the Apify Console. For light usage, set `alert_channel: "dataset_only"` to skip all `spread-alert-triggered` charges and read the dataset directly. For full real-time coverage with Telegram alerts, schedule hourly with a representative ref portfolio. ### How spread detection actually works (read this once) A common question: *"Why do most of my alerts come from Chrono24, with only occasional ones from WatchBox / Bobs / Watchfinder?"* That's the **mathematically expected** behavior — and it's actually the value of cross-platform median: 1. **Cross-platform median = the consensus price.** The actor computes the median price for each sub-ref across every source that returned listings (e.g. Chrono24 + WatchBox + Bobs for `124060`). 2. **Outlier listings below median = arbitrage opportunities.** A spread alert fires when ANY individual listing on ANY platform sits more than `X%` (default 5%) below that consensus. 3. **Where do outliers come from?** Mostly from **Chrono24** because: - Chrono24 has the **highest listing volume** per ref (typically 8-12 listings vs 1-2 on dealer sites) - **More listings = more variance** → statistically more outliers below median - Independent dealers on Chrono24 have **wider pricing dispersion** than US/UK retail dealers (who maintain stable floor prices to protect their own margin) 4. **WatchBox + Bobs + Watchfinder mostly anchor the median.** Their inventory is small but their pricing is **stable and trustworthy** — they're the floor that defines what the watch is *really* worth. They occasionally trigger their own alerts when they get aggressive on a single piece (e.g. the 124060 example above: WatchBox at $10,050 vs the $13,988 cross-platform median). **Practical takeaway**: monitor 15-30 refs (not 3-5) to catch the 1-2 weekly cross-platform spreads on US/UK dealer sites. Chrono24 will dominate alert volume by design — that's where the inefficiency is. The dealer-site signals are rarer but typically larger when they fire. ### Tips - **Use precise references for accurate alerts**: `5711/1A-010` beats `Nautilus`. Sub-models (5711 vs 7011 vs 5990) are tracked separately so a ladies' Nautilus never pollutes the men's median. - **Strict condition filtering**: pair `strict_condition_matching: true` with `platforms: [bobs, watchfinder, europeanwatch]` (the 3 sources that scrape condition reliably) for clean "new only" arbitrage. - **Anti-bot**: Chrono24 and Bobs Watches sit behind Cloudflare. The Actor uses Camoufox + rotating Apify proxy by default — no config needed. - **Watchfinder UK** uses canonical brand+series pages instead of free-text search (which returns no-results for raw refs). The url builder maps known refs to the right grid. - **European Watch Co** lacks reliable site search — the Actor hits the brand grid (`/brand/rolex` etc.) and filters cards by ref-digit substring. - **Watches of Switzerland** hides Patek pricing as "price on request"; those listings are skipped. Rolex/AP listings show full pricing. - **Hodinkee Shop is deprecated** as of 2026-05-04 (acquired by Watches of Switzerland Group, inventory winding down). It's removed from the default `platforms` array but still accepted in the enum for backward-compat scheduled runs. - **EU/US arbitrage**: prices on EU listings are not auto-adjusted in the dataset (you see the raw EUR→USD conversion), but the alert text flags location so you can apply your own landed-cost math. - **Schedule it**: the Actor is designed for hourly cron. Use Apify Schedules → "Run every hour". - **Telegram setup**: `@BotFather` → `/newbot` → token. Add bot to your dealer group → `@userinfobot` in the group → copy chat ID. ### What's new (build 0.1.19 — 2026-05-06) - **Cross-platform median delivering** ✅: build 0.1.18 cloud test confirmed real cross-platform spreads detected on `124060` (Chrono24 + WatchBox + Bobs aggregated) and `116500LN` (Chrono24 + WatchBox aggregated). Top spread example: WatchBox 124060 at $10,050 vs $13,988 cross-platform median = 28.2% below market. - **WatchBox / The 1916 Company URL pattern fix**: previous URL routed to a brand-suggest splash page (0 tiles); switched to `/search/?q=REF` which returns the real product grid. - **Bobs Watches search endpoint fix**: switched from stale `/{brand}-{model}-{page}.html` catalog URLs to `/shop?query=REF` (the actual search endpoint exposed by their homepage form). Cloudflare interstitial timeout extended to 45s. - **Ref filter regression fix**: brand-grid platforms (European Watch Co, Watchfinder, WoS) now match BASE prefix (`5711/1A`) instead of full sub-variant (`5711/1A-010`) so legitimate cards aren't stripped. Aggregator's `extractSubRef` then groups detected sub-variants for accurate median. - **Honest pricing in alerts**: every Telegram message now includes `_Asking price excl. shipping, taxes, customs (varies by buyer location)._` to match what the user sees if they click through to the dealer page. - **Platform names in Telegram**: alerts now show `WatchBox / The 1916 Company` (not the raw `watchbox` enum). Same for all 6 sources. - **Honest dispatcher reporting**: `email_sent` no longer reports `true` after a silent failure. (Bug D — was masking a real problem.) - **Email dropped from MVP** (returning v0.2): Apify's `Actor.call('apify/send-mail')` requires `FULL_PERMISSIONS` scope which the sandboxed runtime token doesn't get for public Actors. Rather than ship a feature that fails silently for end users, email is hidden from the input form. v0.2 will integrate Resend / MailerSend HTTP API directly. Telegram alerts work end-to-end today. ### Earlier — build 0.1.17 (2026-05-05) - **5 critical accuracy bugs fixed**: strict ref-matching per platform, trimmed median (drops outlier listings), 30% min price floor (skip wrong-watch matches), model-name title matching, and **sub-reference grouping** (the big one — distinct sub-models get distinct medians). - **Spending limit respect**: every charge call now honors `ACTOR_MAX_TOTAL_CHARGE_USD`. - **Filter by condition + box & papers** with optional strict mode. - **MCP server (Apify Standby)** for Claude Desktop / Cursor / ChatGPT — see section below. ### FAQ **Is scraping these sites legal?** The Actor only reads publicly-listed prices (no login, no PII). All listings are pages already indexed by Google. Always confirm compliance with each platform's ToS for your specific use case. **Why only Patek/Rolex/AP?** These are the three brands with deep dealer arbitrage — high enough listing volume across multiple platforms to get a meaningful median. We may add Vacheron / Lange / Richard Mille based on demand. **What if a platform changes its DOM?** Each crawler degrades gracefully — a broken selector logs a warning and the run continues with partial data from the other 5 platforms. **Can I add more platforms?** Open an issue and tag your dealer use case. Adding an additional platform is ~30 min of crawler work using the existing template structure. **Disclaimer** Prices are estimates. Always verify the listing manually before any transaction. The Actor is a research tool, not a trading bot — final buy/sell decisions are yours. ### Source notes (history & dealer recognition) If you're new to the secondary luxury watch market, here's the dealer landscape this Actor monitors: - **Chrono24** — the dominant global marketplace (Karlsruhe, Germany; founded 2003). 500K+ active listings; the de-facto reference price for any Patek/Rolex/AP because of pure listing volume. Independent dealers worldwide list here; widest pricing dispersion = where most arbitrage signals come from. - **WatchBox / The 1916 Company** — premium US dealer, originally **WatchBox** (Pennsylvania, founded 2017 by ex-Govberg leadership). In 2024, WatchBox merged with **Govberg Jewelers** (Philadelphia, est. 1916) and **Hyde Park Jewelers** (Denver) to form **The 1916 Company** — the name nods to Govberg's 1916 founding date. The watchbox.com domain redirects to `the1916company.com` since Feb 2025. Dealers still call it "WatchBox" colloquially. Inventory blends ex-WatchBox + ex-Govberg + ex-Hyde Park stock plus new acquisitions. Strong Patek + Rolex; expanding AP. - **Bobs Watches** — Newport Beach CA, est. 1999, Rolex specialist (some Omega/Tudor). Known for transparent buy/sell pricing and long-tenured CEO Paul Altieri. Cloudflare-protected; we use Camoufox + Apify proxy rotation to scrape reliably. - **Watchfinder UK** — UK leader (Maidstone, est. 2002). Acquired by **Richemont Group** in 2018. The cleanest box-and-papers metadata of any source we monitor. Free-text search returns no-results for raw refs, so the Actor maps each ref to its canonical brand+series page. - **European Watch Co (EWC)** — Boston-based independent dealer, est. 2010. Strong Patek inventory (69+ pieces verified 2026-05). Site search is unreliable so we hit the brand grid directly and filter cards by base ref prefix. - **Watches of Switzerland Group (WoS)** — UK + US retail chain group, includes **Mayors** (US), **Mappin & Webb** (UK), **Goldsmiths** (UK), and the WoS-branded boutiques. Patek pricing displayed as "price on request" and is therefore skipped; Rolex/AP work when in stock. - **Hodinkee Shop** — formerly an independent editorial+commerce site (NYC). Acquired by Watches of Switzerland Group in 2024; commerce side wound down through 2026. **Removed from default platforms** list 2026-05-04 but kept in the enum so any pre-existing scheduled runs don't break. Why those 6? They're the largest English-language sources for Patek/Rolex/AP secondary listings — together they represent the bulk of pricing visibility for English-speaking professional dealers (US/UK/CA/AU). German-language Chrono24 covers EU listing flow. Asian-Pacific dealer flow is on Chrono24 + private Telegram groups not yet wired in. ### Use as MCP server (Claude Desktop / Cursor / ChatGPT) In addition to the batch crawler (above), this Actor doubles as a **Model Context Protocol server** in Apify Standby mode. Connect your AI agent and it can query your live arbitrage feed in plain language: *"Show me the biggest Daytona spreads right now"*, *"What's the median price of a 5711/1A this week?"* Three tools are exposed over both REST and MCP JSON-RPC: | Tool | Purpose | |---|---| | `get_arbitrage_snapshot` | Top N current arbitrage opportunities, optionally filtered by ref + min spread % | | `get_market_stats` | Per-ref median, min, max, count across platforms | | `get_listings_by_ref` | Raw listings for a ref, filterable by condition + box/papers, paginated | **Data freshness**: results reflect the most recent batch run (you should schedule the batch crawler to run hourly or however often you need fresh data). If the cache is older than 60 minutes, the server returns 503 with Retry-After. #### Claude Desktop config Add to `claude_desktop_config.json`: ```json { "mcpServers": { "watch-arbitrage": { "url": "https://kazkn--watch-arbitrage-mcp.apify.actor/mcp?token=apify_api_YOUR_TOKEN_HERE", "transport": "streamable-http" } } } ``` Replace `apify_api_YOUR_TOKEN_HERE` with your token from [console.apify.com/account/integrations](https://console.apify.com/account/integrations). #### Cursor / Windsurf Same URL — add it under the MCP section of your IDE settings, transport `streamable-http`. #### Plain HTTP (curl / your backend) ```bash ## Top 10 spreads curl -X POST https://kazkn--watch-arbitrage-mcp.apify.actor/mcp/arbitrage \ -H "Authorization: Bearer apify_api_YOUR_TOKEN" \ -H "Content-Type: application/json" \ -d '{"min_spread_pct": 5, "limit": 10}' ## Market stats for two refs curl -X POST https://kazkn--watch-arbitrage-mcp.apify.actor/mcp/market-stats \ -H "Authorization: Bearer apify_api_YOUR_TOKEN" \ -H "Content-Type: application/json" \ -d '{"references": ["5711/1A", "116500LN"]}' ## Listings: full-set 5711/1A only, first 10 curl -X POST https://kazkn--watch-arbitrage-mcp.apify.actor/mcp/listings \ -H "Authorization: Bearer apify_api_YOUR_TOKEN" \ -H "Content-Type: application/json" \ -d '{"ref": "5711/1A", "box_papers": ["full-set"], "limit": 10}' ``` #### MCP query pricing MCP queries use the same Pay-Per-Event meter as the batch crawler: - `get_arbitrage_snapshot` → $0.50 per call (`spread-alert-triggered`) — you only pay when you ask for the value-extracting view - `get_market_stats` → free read (no event charge) - `get_listings_by_ref` → $0.001 per listing returned (`apify-default-dataset-item`) ### Support - Issues / feature requests: GitHub Issues tab on this Actor - Custom dealer integrations (private feeds, SMS alerts, …): contact via Apify message - See also: [Vinted Smart Scraper](https://apify.com/kazkn/vinted-smart-scraper) — the kazkn portfolio's flagship recurring scraper. # Actor input Schema ## `references` (type: `array`): 🎯 USE PRECISE REFERENCE NUMBERS for accurate arbitrage detection : '5711/1A-010', '116500LN', '124060', '15500ST', '67600ST.OO.1210ST.01'. Mix Patek/Rolex/AP freely (brand auto-detected). 💡 Tip : start with 5-10 references you actually trade — add more later. ⚠️ Model names like 'Nautilus' or 'Daytona' work too but produce LESS PRECISE arbitrage signals (sub-models with different prices get conflated, e.g. men's 5711 vs ladies 7011). Stick to precise references for reliable spread alerts. ## `spread_sensitivity` (type: `string`): Triggers an alert when a listing is BELOW the cross-platform median by at least this %. Lower = more alerts (some noise). Higher = only big opportunities. ## `alert_channel` (type: `string`): Pick your preferred channel. Telegram is fastest for time-sensitive flips. Email digest coming v0.2 (we're verifying delivery permissions on Apify platform first to avoid silently dropped alerts). ## `alert_telegram_bot_token` (type: `string`): 🔧 ONE-TIME setup (5 min): open @BotFather on Telegram → /newbot → name it (e.g. 'My WatchArb Alerts') → copy the token here. Full guide: https://core.telegram.org/bots#how-do-i-create-a-bot ## `alert_telegram_chat_id` (type: `string`): 🔧 To find it: open Telegram → search @userinfobot → tap Start → it tells you your numeric chat ID (e.g. '123456789'). Paste it here. ## `filter_conditions` (type: `array`): Only include listings matching these conditions. Critical for arbitrage : a NEW Patek 5711 vs PRE-OWNED can have $30K+ spread (different market). 🎯 PRECISION GUIDE : Bobs Watches + Watchfinder UK + European Watch scrape condition RELIABLY. Chrono24 + WatchBox + Watches of Switzerland often mark 'unknown' (only on detail page). 💡 TIP : if you want STRICT 'new only' results → (a) uncheck 'unknown' below, AND (b) in Marketplaces section uncheck Chrono24/WatchBox/WoS, OR (c) toggle 'Strict condition matching' ON below. ## `strict_condition_matching` (type: `boolean`): When ON : exclude listings where condition couldn't be reliably scraped from the search page (= 'unknown'). Pairs perfectly with selecting only Bobs + Watchfinder + European Watch in Marketplaces (those 3 scrape condition reliably). When OFF : include 'unknown' listings in your filter results (more volume, less precision). ## `filter_box_papers` (type: `array`): Only include listings with these box/papers status. Critical for arbitrage : a Patek 5711 'full set' commands +20-30% premium vs 'watch only'. Best-effort : Watchfinder + Bobs + European Watch scrape this reliably ; others may mark 'unknown'. Pair with 'Strict condition matching' for clean filtering. ## `platforms` (type: `array`): Default: all 6 active marketplaces. Uncheck any you don't want to monitor (e.g. you only flip US-based, uncheck UK/EU sources). ## `max_listings_per_ref_per_platform` (type: `integer`): 💰 Cost control: caps how many listings to fetch per ref per marketplace. Lower = cheaper run. Default 25 is good for most cases (= 60 refs × 6 platforms × 25 = 9000 max listings/run). ## Actor input object example ```json { "references": [ "5711/1A-010", "5167A-001", "116500LN", "124060", "15500ST.OO.1220ST.04" ], "spread_sensitivity": "5", "alert_channel": "telegram", "filter_conditions": [ "new", "like-new", "very-good", "pre-owned", "unknown" ], "strict_condition_matching": false, "filter_box_papers": [ "full-set", "box-and-papers", "papers-only", "unknown" ], "platforms": [ "chrono24", "watchbox", "bobs", "watchfinder", "europeanwatch", "watchesofswitzerland" ], "max_listings_per_ref_per_platform": 25 } ``` # Actor output Schema ## `arbitrageOpportunities` (type: `string`): Cross-platform spread alerts. Each entry = one listing priced below the cross-platform median for its reference. Sorted by spread % descending (biggest opportunities first). Use this to spot flips fast. ## `marketSnapshot` (type: `string`): Per-reference statistics: median, min, max, listings count, platforms covered. Use this to understand cross-platform pricing baseline before drilling into individual listings. ## `datasetItems` (type: `string`): Every individual listing collected across the 6 marketplaces. Browse, filter, sort by price, condition, year, box and papers. Export as JSON, CSV, Excel, or HTML. # API You can run this Actor programmatically using our API. Below are code examples in JavaScript, Python, and CLI, as well as the OpenAPI specification and MCP server setup. ## JavaScript example ```javascript import { ApifyClient } from 'apify-client'; // Initialize the ApifyClient with your Apify API token // Replace the '' with your token const client = new ApifyClient({ token: '', }); // Prepare Actor input const input = { "references": [ "5711/1A-010", "5167A-001", "116500LN", "124060", "15500ST.OO.1220ST.04" ] }; // Run the Actor and wait for it to finish const run = await client.actor("kazkn/watch-arbitrage-mcp").call(input); // Fetch and print Actor results from the run's dataset (if any) console.log('Results from dataset'); console.log(`💾 Check your data here: https://console.apify.com/storage/datasets/${run.defaultDatasetId}`); const { items } = await client.dataset(run.defaultDatasetId).listItems(); items.forEach((item) => { console.dir(item); }); // 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/js/docs ``` ## Python example ```python from apify_client import ApifyClient # Initialize the ApifyClient with your Apify API token # Replace '' with your token. client = ApifyClient("") # Prepare the Actor input run_input = { "references": [ "5711/1A-010", "5167A-001", "116500LN", "124060", "15500ST.OO.1220ST.04", ] } # Run the Actor and wait for it to finish run = client.actor("kazkn/watch-arbitrage-mcp").call(run_input=run_input) # Fetch and print Actor results from the run's dataset (if there are any) print("💾 Check your data here: https://console.apify.com/storage/datasets/" + run["defaultDatasetId"]) for item in client.dataset(run["defaultDatasetId"]).iterate_items(): print(item) # 📚 Want to learn more 📖? Go to → https://docs.apify.com/api/client/python/docs/quick-start ``` ## CLI example ```bash echo '{ "references": [ "5711/1A-010", "5167A-001", "116500LN", "124060", "15500ST.OO.1220ST.04" ] }' | apify call kazkn/watch-arbitrage-mcp --silent --output-dataset ``` ## MCP server setup ```json { "mcpServers": { "apify": { "command": "npx", "args": [ "mcp-remote", "https://mcp.apify.com/?tools=kazkn/watch-arbitrage-mcp", "--header", "Authorization: Bearer " ] } } } ``` ## OpenAPI specification ```json { "openapi": "3.0.1", "info": { "title": "Watch Arbitrage — Patek/Rolex/AP Cross-Platform Alerts", "description": "Track Patek Philippe, Rolex & Audemars Piguet price spreads across 6 marketplaces (Chrono24, WatchBox, Bobs, Watchfinder UK, European Watch, Watches of Switzerland). Real-time Telegram alerts when a listing is mispriced cross-platform. Filter by condition + box/papers. Built for pro watch dealers.", "version": "0.1", "x-build-id": "ehL9xx5PQmvOC8T8U" }, "servers": [ { "url": "https://api.apify.com/v2" } ], "paths": { "/acts/kazkn~watch-arbitrage-mcp/run-sync-get-dataset-items": { "post": { "operationId": "run-sync-get-dataset-items-kazkn-watch-arbitrage-mcp", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for its completion, and returns Actor's dataset items in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } }, "/acts/kazkn~watch-arbitrage-mcp/runs": { "post": { "operationId": "runs-sync-kazkn-watch-arbitrage-mcp", "x-openai-isConsequential": false, "summary": "Executes an Actor and returns information about the initiated run in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK", "content": { "application/json": { "schema": { "$ref": "#/components/schemas/runsResponseSchema" } } } } } } }, "/acts/kazkn~watch-arbitrage-mcp/run-sync": { "post": { "operationId": "run-sync-kazkn-watch-arbitrage-mcp", "x-openai-isConsequential": false, "summary": "Executes an Actor, waits for completion, and returns the OUTPUT from Key-value store in response.", "tags": [ "Run Actor" ], "requestBody": { "required": true, "content": { "application/json": { "schema": { "$ref": "#/components/schemas/inputSchema" } } } }, "parameters": [ { "name": "token", "in": "query", "required": true, "schema": { "type": "string" }, "description": "Enter your Apify token here" } ], "responses": { "200": { "description": "OK" } } } } }, "components": { "schemas": { "inputSchema": { "type": "object", "required": [ "references", "spread_sensitivity", "alert_channel" ], "properties": { "references": { "title": "1️⃣ Watch references to track (precise refs only — recommended)", "minItems": 1, "type": "array", "description": "🎯 USE PRECISE REFERENCE NUMBERS for accurate arbitrage detection : '5711/1A-010', '116500LN', '124060', '15500ST', '67600ST.OO.1210ST.01'. Mix Patek/Rolex/AP freely (brand auto-detected). 💡 Tip : start with 5-10 references you actually trade — add more later. ⚠️ Model names like 'Nautilus' or 'Daytona' work too but produce LESS PRECISE arbitrage signals (sub-models with different prices get conflated, e.g. men's 5711 vs ladies 7011). Stick to precise references for reliable spread alerts.", "items": { "type": "string" } }, "spread_sensitivity": { "title": "2️⃣ How aggressive should the alerts be?", "enum": [ "3", "5", "8", "10", "15", "20" ], "type": "string", "description": "Triggers an alert when a listing is BELOW the cross-platform median by at least this %. Lower = more alerts (some noise). Higher = only big opportunities.", "default": "5" }, "alert_channel": { "title": "3️⃣ How do you want to get alerts?", "enum": [ "telegram", "dataset_only" ], "type": "string", "description": "Pick your preferred channel. Telegram is fastest for time-sensitive flips. Email digest coming v0.2 (we're verifying delivery permissions on Apify platform first to avoid silently dropped alerts).", "default": "telegram" }, "alert_telegram_bot_token": { "title": "Telegram bot token", "type": "string", "description": "🔧 ONE-TIME setup (5 min): open @BotFather on Telegram → /newbot → name it (e.g. 'My WatchArb Alerts') → copy the token here. Full guide: https://core.telegram.org/bots#how-do-i-create-a-bot" }, "alert_telegram_chat_id": { "title": "Your Telegram chat ID", "type": "string", "description": "🔧 To find it: open Telegram → search @userinfobot → tap Start → it tells you your numeric chat ID (e.g. '123456789'). Paste it here." }, "filter_conditions": { "title": "Filter by condition", "uniqueItems": true, "type": "array", "description": "Only include listings matching these conditions. Critical for arbitrage : a NEW Patek 5711 vs PRE-OWNED can have $30K+ spread (different market). 🎯 PRECISION GUIDE : Bobs Watches + Watchfinder UK + European Watch scrape condition RELIABLY. Chrono24 + WatchBox + Watches of Switzerland often mark 'unknown' (only on detail page). 💡 TIP : if you want STRICT 'new only' results → (a) uncheck 'unknown' below, AND (b) in Marketplaces section uncheck Chrono24/WatchBox/WoS, OR (c) toggle 'Strict condition matching' ON below.", "items": { "type": "string", "enum": [ "new", "like-new", "very-good", "good", "fair", "vintage", "pre-owned", "unknown" ], "enumTitles": [ "🆕 New / Unworn (factory-fresh, full set)", "✨ Like new / Mint (worn briefly, no signs of wear)", "👍 Very good (light wear, fully serviced)", "✅ Good (visible wear, normal use)", "⚠️ Fair (notable wear, may need service)", "🕰️ Vintage (pre-2000, collector grade)", "📦 Pre-owned (generic — used by some marketplaces as catch-all)", "❓ Unknown (condition not specified on listing — include for completeness)" ] }, "default": [ "new", "like-new", "very-good", "pre-owned", "unknown" ] }, "strict_condition_matching": { "title": "🔒 Strict condition matching (recommended for 'new only' arbitrage)", "type": "boolean", "description": "When ON : exclude listings where condition couldn't be reliably scraped from the search page (= 'unknown'). Pairs perfectly with selecting only Bobs + Watchfinder + European Watch in Marketplaces (those 3 scrape condition reliably). When OFF : include 'unknown' listings in your filter results (more volume, less precision).", "default": false }, "filter_box_papers": { "title": "Filter by box & papers", "uniqueItems": true, "type": "array", "description": "Only include listings with these box/papers status. Critical for arbitrage : a Patek 5711 'full set' commands +20-30% premium vs 'watch only'. Best-effort : Watchfinder + Bobs + European Watch scrape this reliably ; others may mark 'unknown'. Pair with 'Strict condition matching' for clean filtering.", "items": { "type": "string", "enum": [ "full-set", "box-and-papers", "papers-only", "box-only", "watch-only", "unknown" ], "enumTitles": [ "🏆 Full set (box + papers + tags + extras)", "📦📄 Box + papers (no tags/extras)", "📄 Papers only (no box)", "📦 Box only (no papers)", "🕰️ Watch only (no box, no papers)", "❓ Unknown (not specified — include for completeness)" ] }, "default": [ "full-set", "box-and-papers", "papers-only", "unknown" ] }, "platforms": { "title": "Marketplaces to monitor", "uniqueItems": true, "type": "array", "description": "Default: all 6 active marketplaces. Uncheck any you don't want to monitor (e.g. you only flip US-based, uncheck UK/EU sources).", "items": { "type": "string", "enum": [ "chrono24", "watchbox", "bobs", "watchfinder", "europeanwatch", "watchesofswitzerland", "hodinkee" ], "enumTitles": [ "🌍 Chrono24 (largest global marketplace)", "🇺🇸 WatchBox / The 1916 Company (premium dealer, USA)", "🇺🇸 Bobs Watches (Rolex specialist, USA)", "🇬🇧 Watchfinder UK (UK leader)", "🇺🇸 European Watch Company (Boston dealer)", "🇬🇧 Watches of Switzerland (UK+US retail chain)", "⚠️ Hodinkee Shop (deprecated — winding down 2026, kept for legacy)" ] }, "default": [ "chrono24", "watchbox", "bobs", "watchfinder", "europeanwatch", "watchesofswitzerland" ] }, "max_listings_per_ref_per_platform": { "title": "Max listings per reference per platform", "minimum": 1, "maximum": 200, "type": "integer", "description": "💰 Cost control: caps how many listings to fetch per ref per marketplace. Lower = cheaper run. Default 25 is good for most cases (= 60 refs × 6 platforms × 25 = 9000 max listings/run).", "default": 25 } } }, "runsResponseSchema": { "type": "object", "properties": { "data": { "type": "object", "properties": { "id": { "type": "string" }, "actId": { "type": "string" }, "userId": { "type": "string" }, "startedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "finishedAt": { "type": "string", "format": "date-time", "example": "2025-01-08T00:00:00.000Z" }, "status": { "type": "string", "example": "READY" }, "meta": { "type": "object", "properties": { "origin": { "type": "string", "example": "API" }, "userAgent": { "type": "string" } } }, "stats": { "type": "object", "properties": { "inputBodyLen": { "type": "integer", "example": 2000 }, "rebootCount": { "type": "integer", "example": 0 }, "restartCount": { "type": "integer", "example": 0 }, "resurrectCount": { "type": "integer", "example": 0 }, "computeUnits": { "type": "integer", "example": 0 } } }, "options": { "type": "object", "properties": { "build": { "type": "string", "example": "latest" }, "timeoutSecs": { "type": "integer", "example": 300 }, "memoryMbytes": { "type": "integer", "example": 1024 }, "diskMbytes": { "type": "integer", "example": 2048 } } }, "buildId": { "type": "string" }, "defaultKeyValueStoreId": { "type": "string" }, "defaultDatasetId": { "type": "string" }, "defaultRequestQueueId": { "type": "string" }, "buildNumber": { "type": "string", "example": "1.0.0" }, "containerUrl": { "type": "string" }, "usage": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "integer", "example": 1 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 } } }, "usageTotalUsd": { "type": "number", "example": 0.00005 }, "usageUsd": { "type": "object", "properties": { "ACTOR_COMPUTE_UNITS": { "type": "integer", "example": 0 }, "DATASET_READS": { "type": "integer", "example": 0 }, "DATASET_WRITES": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_READS": { "type": "integer", "example": 0 }, "KEY_VALUE_STORE_WRITES": { "type": "number", "example": 0.00005 }, "KEY_VALUE_STORE_LISTS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_READS": { "type": "integer", "example": 0 }, "REQUEST_QUEUE_WRITES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_INTERNAL_GBYTES": { "type": "integer", "example": 0 }, "DATA_TRANSFER_EXTERNAL_GBYTES": { "type": "integer", "example": 0 }, "PROXY_RESIDENTIAL_TRANSFER_GBYTES": { "type": "integer", "example": 0 }, "PROXY_SERPS": { "type": "integer", "example": 0 } } } } } } } } } } ```