{
  "markdown": "# @pipeworx/phillips\n\nRealized auction prices from Phillips — look up an artist or maker and get\ntheir recent past lots with the price actually paid (hammer plus buyer's\npremium), the estimate range it sold against, currency, lot and sale number,\nsaleroom, department and sale date; pull full detail for any single lot.\n\nPart of [Pipeworx](https://pipeworx.io) — an MCP gateway connecting AI agents to 1546+ live data sources.\n\n## Tools\n\n- `phillips_results_search(artist?, artist_id?, category?, include_upcoming?, limit?)`\n  — realized prices for one maker. Resolves the name against Phillips' own\n  artist index, then reads that maker's artist page. `category` is a free-text\n  filter matched against department, saleroom, sale number and lot title (e.g.\n  \"jewels\", \"watches\", \"editions\", \"photographs\", \"design\", \"London\"). Pass\n  `artist_id` from a previous result to skip name resolution entirely.\n- `phillips_lot_details(lot_id, slug?)` — full detail for one lot, keyed by the\n  `lot_id` a `phillips_results_search` result carries. Adds the hammer price\n  (before premium) and the full sale name.\n\nTool names are house-prefixed on purpose. Bonhams shipped first with bare\n`results_search`/`lot_details`; with three art houses live those two names\nwould collide across packs and `ask_pipeworx` routing would have nothing to\ndisambiguate on. New house packs should keep the `{house}_` prefix.\n\n## Auth\n\nKeyless. Public pages, no login required.\n\n## Data sources\n\nPhillips' **keyword search is closed to crawlers** — robots.txt (checked\n2026-09-05) disallows `/Search`, `/search` AND `/SEARCH`, so all three case\nvariants of the obvious path are out, along with `/bin/`, `/phillips/otis` and\n`/*/filter/`. This pack therefore does not search at all; it goes by maker,\nthrough three paths robots.txt leaves open:\n\n- `https://www.phillips.com/sitemap.xml` — 13,031 `/artist/{makerId}/{name}`\n  URLs. This is the artist-name → makerId index names are resolved against.\n  Fetched at most once per isolate. **The names in it are percent-encoded AND\n  HTML-escaped** (`Jennifer%20&%20Kevin%20McCoy`); take both layers off\n  before you use one, or the URL you build goes out double-encoded and 404s.\n- `https://www.phillips.com/artist/{makerId}/{slug}` — the artist landing page.\n  Realized prices are server-rendered here, and the same rows also ship as JSON\n  in the `React.createElement(PhillipsReact.ArtistLanding, {...})` hydration\n  props: the `maker` prop is a JSON *string* holding\n  `pastLots {currentPage, totalPages, totalCount, resultsPerPage, data[]}`.\n  Each row carries `hammerPlusBP` (the realized price), `lowEstimate`,\n  `highEstimate`, `currencySign`, `saleNumber`, `lotNumber`, `objectNumber`\n  (the lot id) and `auctionStartDateTimeOffset`. **`hammerPlusBP` is `0` for a\n  lot that did not sell**, which is an absent price, not a price of zero.\n- `https://www.phillips.com/detail/{slug}/{objectNumber}` — the lot page.\n\nThe `{slug}` segment of both `/artist/` and `/detail/` URLs is decorative — any\nnon-empty value resolves the same record, only the numeric id is read.\n\n### Coverage ceiling, and why it is there\n\nThe artist page serves **one page of past lots — 24 rows**. Page 2 onward is\nfetched by the browser from a different host (`api.phillips.com`), and there is\nno query parameter that pages the HTML page itself (`?page=2`, `?p=2` and\n`?pastLotsPage=2` were each tried and all return page 1). So\n`phillips_results_search` returns the 24 most recent past lots and reports\n`total_past_lots` alongside them, so a caller can see how much archive it is\nnot seeing — Banksy returns 24 of 272, Patek Philippe 24 of 3,728.\n\n### Reading the lot page\n\nThe lot page is a different stack from the rest of the site (Remix; the state\nblob is a flattened key-then-value array, not an addressable object), so this\npack reads it from three places and prefers the rendered HTML:\n\n- **JSON-LD `Product`** — title, maker, image, currency, sold/unsold. Note\n  `offers.price` is the **low estimate**, not a realized price; do not treat it\n  as one.\n- **Rendered HTML** — `Sold For {cur}{amount}` and `Estimate {cur}{low}–{high}`.\n  The estimate MUST come from here: the only `lowEstimate`/`highEstimate` pair\n  in the state blob belongs to the page's currency-conversion table, so reading\n  it returns a plausible number in the wrong currency (HK$622,000–830,000 for a\n  lot whose real estimate is £60,000–80,000).\n- **State blob**, for the two fields nothing else carries — `hammerPrice` and\n  `auctionStartDateTime`. Because key-to-value adjacency in a flattened array\n  is not guaranteed, the hammer price is dropped rather than returned if it\n  exceeds the realized price.\n\n### Sale numbers encode department and saleroom\n\nA sale number is `{2-letter saleroom}{2-digit department}{2-digit sequence}{2-digit year}`\n— `UK010526` is London, department 01, sale 05, 2026. The department mapping\nbelow was derived from, and agreed with, all 892 past auctions listed on\n`/auctions/past` (2026-09-05): `01` Modern & Contemporary Art, `03` Editions &\nWorks on Paper, `04` Photographs, `05` Design, `06` Jewels, `07` Editions,\nPhotographs and Design, `08` Watches. `00` and `09` are **not** departments —\nthey are sale formats (special/collaboration, and online) that run across\ndepartments — and are labelled as such rather than guessed at. Salerooms seen:\n`NY` New York, `UK` London, `HK` Hong Kong, `CH`/`GE` Geneva.\n\n## Related packs\n\n`@pipeworx/bonhams` and `@pipeworx/sothebys` cover the same question at the\nother two houses. Christie's is out of scope (closed, internal API).\n\n## Quick Start\n\nAdd to your MCP client (Claude Desktop, Cursor, Windsurf, etc.):\n\n```json\n{\n  \"mcpServers\": {\n    \"phillips\": {\n      \"url\": \"https://gateway.pipeworx.io/phillips/mcp\"\n    }\n  }\n}\n```\n\n### What this endpoint actually serves\n\n`tools/list` at `https://gateway.pipeworx.io/phillips/mcp` returns the tools in the table\nabove **plus the shared Pipeworx meta-tools** — `ask_pipeworx`,\n`discover_tools`, `search_within`, `remember`/`recall` and the rest of the\ngateway-wide set. So the tool count you see is larger than this table: a\nsingle-pack endpoint currently lists roughly 30 shared tools alongside the\npack's own. The connection's `initialize` response states its exact scope, and\nis the authoritative answer for a given day.\n\nThis is deliberate, not multiplexing by accident. The meta-tools are what let a\nscoped connection answer a question this pack does not cover — via\n`ask_pipeworx`, which routes across the whole catalog — without you adding a\nsecond MCP server. There is currently no way to mount a pack endpoint without\nthem; if the extra schemas cost you more context than the routing is worth,\nconnect to the full gateway once rather than to several pack endpoints.\n\nOr connect to the full Pipeworx gateway to get every pack's tools listed\ndirectly, instead of just this one's:\n\n```json\n{\n  \"mcpServers\": {\n    \"pipeworx\": {\n      \"url\": \"https://gateway.pipeworx.io/mcp\"\n    }\n  }\n}\n```\n\nBoth URLs reach the same gateway and the same 1546+ data sources. The\nonly difference is which pack's tools are listed **directly**; `ask_pipeworx`\nreaches all of them from either one.\n\n## Using with ask_pipeworx\n\nInstead of calling tools directly, you can ask questions in plain English —\nthis works on the pack endpoint above as well as on the full gateway:\n\n```\nask_pipeworx({ question: \"your question about Phillips data\" })\n```\n\nThe gateway picks the right tool and fills the arguments automatically.\n\n## More\n\n- [Docs and guides](https://pipeworx.io/docs)\n- [pipeworx.io](https://pipeworx.io)\n\n## License\n\nMIT\n",
  "bytes": 7624,
  "sha": "21b63659a6ffb0b2e3dab43b588cf8472929971f3c135025dc40c8d5861c79e1",
  "repo_slug": "pipeworx-io/mcp-phillips",
  "fonte": "repo",
  "truncated": false,
  "api": "https://agentalog.com/api/listings/mcp_io_github_pipeworx_io_phillips_25ef2851/readme"
}