WineAuction

io.github.antyserPublished by GitHub account antyserv1.2.1Updated Nov 3, 2026

Wine prices: retail, auction and restaurant references; check any quoted price.

VerifiedStreamable HTTPWeb executableLocation & Lifestyle

Overview

WineAuction's hosted MCP server (wineauction.ai); OAuth sign-in required; Streamable HTTP.

No AI overview yet: these facts come from the registry entry and the server itself.

Endpoint
wineauction.ai
Last checked

Installation

In SourceWeft

  1. Open WineAuction in the dashboard and add it to a workspace.
  2. Connect the required credentials. They are stored privately and never shown on public pages.
  3. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.antyser%2Fwineauction/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.

{
  "mcpServers": {
    "wineauction": {
      "type": "http",
      "url": "https://wineauction.ai/api/mcp"
    }
  }
}

README

WineAuction plugin

Compare presented wine offers with retail and auction market prices, and find wine prices, historical auction sales, current offers, and available critic reviews. The data comes from WineAuction.

What it adds

  • WineAuction connector (https://wineauction.ai/api/mcp): you sign in with your WineAuction account the first time you use it.
  • Skills, one per use case:
    • restaurant-wine-list: a restaurant wine list you bring (photo, PDF, link or text), or a named restaurant. Shows the markup of each bottle over retail and auction prices.
    • merchant-offer-check: a merchant email, newsletter, futures offer, price list or a single quoted price. Judges each price against the market.
    • auction-lots: an auction sheet or lots you bring, or a named auction or house. Lots below market, lots ending soon, sale results, and a recurring watch.
    • wine-info: a wine, producer, vintage or region: retail and auction prices, past sales, critic reviews and where it is currently offered.
    • cellar-valuation: a list of bottles you own. Value per bottle and in total, at auction and at retail.
    • wineauction-reference: shared tool fields, price-reading rules, list parsing and the HTML report renderer that the skills above link to.

Version 1.2.1 explains the auction reference's fallback: the exact vintage when it has sales, otherwise all vintages pooled, with vintage_basis saying which.

Install

Claude Code:

/plugin marketplace add antyser/wineauction-plugin/plugin install wineauction@wineauction

Then run /mcp and sign in to wineauction.

Codex:

codex plugin marketplace add antyser/wineauction-plugin

Then install WineAuction from the plugin directory and sign in when asked. The repo carries both manifests: .claude-plugin/ and .mcp.json for Claude, plugin.json, mcp.json and .agents/plugins/ for OpenAI. The skills in skills/ are shared.

Claude (web, desktop and Cowork), connector only: open Customize > Connectors, search for WineAuction in the directory, click Connect and sign in with your WineAuction account.

Claude with the skills: go to Customize > Plugins, add the marketplace antyser/wineauction-plugin, then install WineAuction. Open the plugin's Connectors tab and connect the WineAuction connector, signing in with your WineAuction account.

ChatGPT (connector only, until the plugin is listed in ChatGPT's directory): open chatgpt.com/plugins, click Add, then Add custom MCP server. Name it WineAuction, paste https://wineauction.ai/api/mcp, keep Authentication on OAuth, tick "I understand and want to continue", click Create as a plugin, and sign in with your WineAuction account.

Step-by-step instructions for every app are at wineauction.ai/mcp.

Data

The plugin does not collect anything itself. Each tool call sends your query (wine names, filters, or the prices you ask it to evaluate) to https://wineauction.ai/api/mcp and returns results. All tools are read-only. The HTML report is generated locally and loads Google Fonts when viewed with network access; without it, system fonts are used. See the privacy policy and terms.

Support

Open an issue at github.com/antyser/wineauction-plugin/issues.

Notes

  • Prices are USD per 750 ml. Auction prices include the buyer's premium.
  • evaluate_offers judges up to 25 rows per call in the skills; long lists are sent in batches.
  • The tools only read data. They cannot bid, buy or contact anyone.

Build and verify a release

Both versioned manifests and server.json use the same release version. server.json is the official MCP Registry entry; .github/workflows/mcp-registry.yml publishes it when it changes on main, and the registry only accepts a new version. The portable package keeps plugin.json, mcp.json, skills/ and assets/ at the repository root; the Claude marketplace manifests remain available alongside them. The 30-row newsletter under review/fixtures/ is synthetic supplied-price input for workflow review, not a set of real merchant offers or market references.

Natural-language skill-routing examples are kept separately in review/trigger-cases.json, one per task skill; the OpenAI review cases continue to cover the six connector tools and the existing workflow examples.

After committing the release candidate, build the ZIP from tracked files with:

python3 scripts/build_release.py

The default output is ~/Downloads/wineauction-plugin-1.2.1.zip. To verify a candidate commit and its packaged renderer, run:

scripts/local_ci.sh "$(git rev-parse HEAD)"

The local check requires a clean checkout at the supplied commit, validates the manifests and review cases, builds and inspects the ZIP, and exercises the renderer from the archive. It uses Python 3 and Node.js.

OpenAI package

OpenAI's plugin review blocks two skills, so the ZIP uploaded to OpenAI is a variant of the full plugin, built from the same commit:

  • python3 scripts/build_release.py builds the full plugin for Claude and Codex (all five task skills and skills/wineauction-reference).
  • python3 scripts/build_openai_package.py builds ~/Downloads/wineauction-plugin-1.2.1-openai.zip. It drops skills/auction-lots, moves skills/wineauction-reference to a root shared/ folder without its SKILL.md and agents/, rewrites the ../wineauction-reference/ links to ../../shared/, keeps only plugin.json, mcp.json, LICENSE, assets/, skills/ and shared/, and applies the files under openai/ over the result (same paths as in the package): the app-name plugin.json, an mcp.json without its own OAuth block, an OpenAI-specific README.md, the four task SKILL.md files and the shared references that use neutral price-research wording. The full plugin ZIP also carries openai/; the repo has no .gitattributes export rules because the Claude directory rejects them.

The OpenAI build checks that no file links to a missing path, that nothing still references auction-lots or wineauction-reference, and that no secret-like material is packaged. Both builds run in scripts/local_ci.sh. Edit the openai/ copy, not the tracked file, to change what OpenAI receives.

Source: README.md at commit b0c57c5

Tools

0
This server asks you to sign in, so SourceWeft could not read its tools without an account and has not indexed them. What a connected account sees can depend on that account and its permissions.

Version history

1
  1. v1.2.1LatestNov 3, 2026

Score

85% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts