Vivino MCP Server - AI Wine Recommendations

Vivino MCP Server connects AI assistants — Claude Desktop, ChatGPT, and any MCP-compatible agent — to live Vivino wine data via the Model Context Protocol. Let an LLM search wines, compare ratings and prices, inspect regions and styles, and recommend bottles for a meal, budget, occasion or cellar shortlist.

Ask Claude: "Find a white wine for shellfish under $50, using Vivino ratings and prices, and explain why you picked it." The server calls Vivino data tools and returns ranked recommendations with rating, price, region, grape/style, food-pairing signals and caveats.

No Vivino account, no browser automation, no coding. Connect your AI assistant and start asking wine questions in plain language.

What is Vivino MCP Server?

Vivino MCP Server runs in Apify Standby mode as a persistent MCP server. Conversational AI assistants call it over Streamable HTTP at /mcp to retrieve and reason over Vivino wine data.

It is designed as the AI-facing companion to MrBridge's Vivino scrapers:

• Vivino Wine Scraper discovers wines by region, producer, grape, type, price, rating and vintage.
• Vivino Wine Data Scraper enriches known wines from names or Vivino URLs with ratings, prices, taste profiles, reviews and details.

This MCP server focuses on live LLM workflows: quick lookup, filtered exploration, wine details and recommendation reasoning.

• Recommendation answers grounded in Vivino data — ratings, ratings count, price, origin, style, grapes, image, URL and food-pairing signals.
• Meal and budget aware — ask for Champagne alternatives, Chablis for shellfish, steak reds under $40, or highly rated Burgundy whites.
• Search by text or structured filters — query by wine name, appellation, producer, style or use filters for rating, price, country, grape, type and food IDs.
• Currency-aware pricing — choose one of five supported currencies for Vivino prices.
• LLM-ready normalized output — compact JSON designed for Claude, ChatGPT, agents and RAG workflows.
• Apify-hosted MCP endpoint — no local server to run; connect through Apify token auth.

Quick Start - Connect in 2 minutes

Your connection URL is the same for every MCP client — replace YOUR_APIFY_TOKEN with your Apify API token:

https://mrbridge--vivino-mcp-server.apify.actor/mcp?token=YOUR_APIFY_TOKEN

1. Connect your AI assistant

• Claude Desktop — Settings -> Connectors -> Add custom connector, paste the URL above, save.
• ChatGPT (Plus / Pro / Team / Enterprise) — Settings -> Connectors, enable developer mode, Add custom connector, paste the URL, save.
• Apify Universal MCP — already using mcp.apify.com? Add this server as a tool: https://mcp.apify.com?tools=mrbridge/vivino-mcp-server.

2. Start asking questions

Try:

"Recommend 3 white wines for oysters under $50. Prefer France, explain rating, price and food-pairing evidence."

The endpoint speaks Streamable HTTP at /mcp.

Real Example: AI wine recommendation from Vivino data

A user asks:

"Recommend a white wine for shellfish under 50 USD. I like Chablis."

The assistant calls recommend_wines with searchQuery: "Chablis", foods: ["shellfish"], preferredStyles: ["white"], and maxBudget: 50.

A typical result:

Wine	Rating	Price	Why it fits
Daniel-Etienne Defaix Chablis Premier Cru 'Vaillon' 2015	4.2	$49.44	Burgundy Chablis, Chardonnay, shellfish pairing, under budget
La Chablisienne Chablis 1er Cru 'L'Homme Mort' 2022	4.2	$48.43	White Burgundy style, shellfish pairing, strong rating count

The assistant can then explain the tradeoff: both fit the budget and pairing; choose Defaix for a mature 2015 Premier Cru profile, or La Chablisienne for a younger Chablis with similar score and slightly lower price.

Every recommendation is derived from the wine records returned by Vivino-facing tools, not from a generic wine-pairing list.

What tools are available? (4 tools)

Wine search and exploration

Tool	What it does
search_wines	Search Vivino by free text: wine, producer, appellation, grape, region or style. Uses Vivino exploration search and returns normalized wine records.
explore_wines	Explore wines with structured filters: country codes, grape IDs, wine type IDs, food IDs, rating, price, sort and pagination.

Details and recommendations

Tool	What it does
get_wine_details	Fetch richer details for a Vivino wine or vintage by wineId, vintageId or Vivino URL.
recommend_wines	Rank wines for an LLM user request using Vivino data, budget, food, style and country preferences. Returns fit score, reasons and caveats.

What data can AI use?

The MCP normalizes the same families of data exposed by the MrBridge Vivino scrapers:

Data	Used by the MCP
Wine identity	Wine name, winery, vintage/year, Vivino IDs and URL
Ratings	Average rating, ratings count and ratingScope (vintage, wine, unknown) when Vivino exposes the distinction
Market data	Price, currency and market availability when Vivino returns a current listing
Origin	Region, country, appellation/style signals
Wine style	Red, white, sparkling, rose, dessert, fortified; style names such as Burgundy Chablis
Grapes	Grape varieties when present, with grapesReliability; appellation/style grapes should be treated as directional, not producer-specific truth
Food pairing	Vivino style food pairings such as shellfish, pork, rich fish, poultry, beef, lamb, cheese
Media	Bottle or label image URL when available
Taste structure	Vivino body, acidity, tannin, sweetness, intensity, fizziness, style descriptions and taste.reliability when present
Search match	searchMatch shows whether a free-text query matched exactly, partially or weakly, including missing producer/appellation terms
Caveats	Missing price, unavailable rating, wine-level rating aggregation, generic grapes, unavailable taste structure, above-budget result, inferred food fit

For large batch extraction, scheduled market monitoring, reviews, full taste profiles or 250+ named wine enrichment, use the dedicated scrapers. For interactive AI recommendations, use this MCP server.

What can AI do with Vivino data?

• Meal pairing — "Find a red for lamb under EUR 35 with at least 4.0 Vivino rating."
• Restaurant shortlist — "Build a 10-bottle by-the-glass shortlist: 3 sparkling, 3 whites, 3 reds, 1 dessert."
• Retail discovery — "Show highly rated Italian reds under $30 with enough ratings to trust the score."
• Cellar planning — "Compare Chablis, Sancerre and white Burgundy options for a seafood dinner."
• Gift selection — "Recommend a Champagne-style sparkling wine under $60 for a birthday."
• Agent workflows — pair with the full Vivino scrapers to enrich a catalog, then use this MCP for conversational selection.

Example prompts

• "Search Vivino for Opus One and summarize the top 3 results with ratings and prices."
• "Explore French white wines rated 4.1+ under 40 USD and recommend the safest pick."
• "Recommend 5 sparkling wines for aperitif under EUR 30. Mention caveats if price is missing."
• "I have roast chicken and soft cheese. Find wines that match both and explain why."
• "Get details for this Vivino URL and turn it into a product-card summary."

Configuration

The server works in English by default. Optional Actor input fields tune currency, response size and timeout:

Field	Type	Default	Description
currencyCode	select	USD	Currency used for price data: USD, EUR, CNY, JPY or KRW. The Vivino market is derived automatically from the selected currency.
maxToolResults	integer	10	Maximum results any MCP tool can return.
requestTimeoutSecs	integer	20	Timeout for upstream Vivino HTTP requests.

Standby mode and performance

The server runs in Apify Standby mode as a hosted HTTP MCP server.

• Warm requests: fast enough for conversational use once the Actor is running.
• Cold start: the first request after inactivity may take longer while Apify starts the container.
• Upstream dependency: Vivino can rate-limit or change internal endpoints. The server returns clear tool errors when upstream data is unavailable.

How much does it cost?

This Actor is designed for Apify pay-per-event monetization. Check the Pricing tab for the live configuration on your Apify plan.

Recommended pay-per-event setup in Apify Console:

Event name	What it bills
tool-read	Read operations such as search_wines and explore_wines
tool-action	Higher-value operations such as get_wine_details and recommend_wines
apify-actor-start	Actor startup at Apify's default low price

Do not enable apify-default-dataset-item; tool calls already record dataset rows for observability, and charging both the tool event and the dataset item would double bill the same user action.

For high-volume extraction priced per delivered wine row, use:

• Vivino Wine Scraper for regional discovery.
• Vivino Wine Data Scraper for name/URL enrichment.

When should I use this MCP vs the scrapers?

Need	Best tool
Ask Claude or ChatGPT for wine recommendations in conversation	Vivino MCP Server
Search a few wines and reason over fit, budget and food pairing	Vivino MCP Server
Extract hundreds or thousands of wines by region, grape, producer or type	Vivino Wine Scraper
Enrich a known list of wine names or Vivino URLs with reviews/taste profiles	Vivino Wine Data Scraper
Build scheduled market monitoring datasets	The scrapers, then optionally query summaries through an LLM

Apify Universal MCP and other clients

Already using the Apify Universal MCP? Include this server as a tool:

https://mcp.apify.com?tools=mrbridge/vivino-mcp-server

Any MCP-compatible client can call the same Streamable HTTP endpoint at /mcp with your Apify token in the Authorization: Bearer header or ?token= query parameter.

Integrate with Zapier, Make and n8n

Vivino MCP Server runs on Apify, so it can be invoked from automation platforms via Actor ID sviYkC8LecqIBeb9N using the standard Apify integrations for Zapier, Make and n8n.

For real-time AI conversations, use the MCP connection URL. For batch jobs, call the Actor through the Apify API or use the dedicated Vivino scrapers.

FAQ

Does Vivino have an official API?

No public API is available for these use cases. This server uses publicly reachable Vivino endpoints and normalizes their responses for LLM use. Upstream endpoint changes can require maintenance.

Do I need a Vivino account?

No. The MCP server does not require a Vivino login or browser session.

Does this return reviews and full taste profiles?

The MCP focuses on compact live recommendation data. For full reviews and detailed taste profiles, use Vivino Wine Data Scraper. For taste profiles across regional discovery runs, use Vivino Wine Scraper.

Why is price sometimes missing?

Vivino price availability depends on country, currency, merchant listings and the wine itself. Missing price is returned as a caveat; the assistant should not claim availability unless a current price or URL is present. Grapes may be generic appellation/style metadata, and ratings can be aggregated at wine level when the same counter repeats across multiple fetched vintages; inspect grapesReliability, ratingScope, taste.reliability, searchMatch and dataWarnings. If searchMatch.level is partial or weak, say which requested terms were not matched before using the result.

Can I use it with ChatGPT?

Yes. Use ChatGPT connectors with developer mode and add the MCP URL from the Quick Start section. Claude Desktop and other MCP clients work the same way.

Is this affiliated with Vivino?

No. This Actor is built by MrBridge and is not affiliated with or endorsed by Vivino.

Other wine-data Actors by MrBridge

Actor	Best for
Vivino Wine Scraper	Discover wines by region, grape, producer, rating, price and vintage
Vivino Wine Data Scraper	Enrich known wine names or Vivino URLs with ratings, prices, reviews and taste profiles
Millesima Wine Scraper	Wine prices and critic ratings from major reviewers
Wine-Searcher Scraper from List	Popularity, critic scores and prices from known wine lists

Built by MrBridge

This MCP server is built by MrBridge. Explore more wine data tools and MCP servers at mr-bridge.com and the MrBridge MCP server catalog.

Resources

• Apify MCP documentation
• Model Context Protocol
• Vivino Wine Scraper
• Vivino Wine Data Scraper

Legal and responsible use

This Actor retrieves publicly available wine information from Vivino-facing endpoints. You are responsible for ensuring your use complies with Vivino's Terms of Service and applicable law. Do not use the output to misrepresent wine availability, pricing or endorsements. Vivino MCP Server is not endorsed by and is not affiliated with Vivino.