Liquid Grape x Wine Labs

An AI-Native Workflow for Fine Wine

Liquid Grape is building a modern home for fine wine buyers and collectors. Henrik and the team wire AI agents directly into the Wine Labs MCP, turning a single prompt into the kind of vintage-by-vintage analysis that used to take an analyst a day.

About Liquid Grape

Liquid Grape is a fine wine platform led by Henrik Maaß and Annie Dörfelt, focused on transparent pricing and serious collector tooling. The team uses Wine Labs daily to value inventory, follow the market across exchanges and auction houses, and feed AI agents that generate bespoke client analyses.

Visit liquidgrape.de to see what they are building.

Industry Reality

A credible fine wine recommendation pulls together pricing, vintage performance and critic scores across many sources. Doing that at the speed of a modern merchant needs real infrastructure underneath.

• Per-bottle research took hours: price history, format, scores and value lived in different tools, so every client question meant another manual pull.
• Cross-source prices are hard to line up: exchanges, auction houses and retailers each speak their own dialect, and comparing them cleanly takes careful plumbing.
• AI agents need a real data spine: agents reason over live market data only if that data is clean, normalized and exposed through an interface they can call.

What We Delivered

Wine Labs MCP, plugged straight into their agents

The Wine Labs MCP exposes prices, market signals and critic scores to any MCP-aware client. Henrik connected Perplexity Computer with a single key and was querying the full surface immediately.

The chart below is a real artifact. Henrik prompted the agent for a vintage-by-vintage analysis of a Meursault 1er Cru; the agent pulled data through the MCP and rendered it in Liquid Grape's house style, in German. No spreadsheet work.

Exchanges, auctions and retail side-by-side

Day to day, the team uses Wine Labs to compare prices across the three channels that matter for fine wine on a single timeline. Spotting divergence between sources is now a primary workflow.

Private offers that sit inside the market

Liquid Grape uploads allocations and private deals through the Wine Labs custom feed. Those offers then appear next to the rest of the market in price history, comparisons and MCP responses.

Impact

• Full vintage analyses in minutes: "which vintage is best value right now?" becomes a single prompt, with a finished, client-shareable chart at the end.
• One answer across channels: with exchanges, auction houses and retailers unified in one place, the team can value a bottle from every angle at once.
• Their book, priced in context: private offers are benchmarked against the wider market automatically, surfacing mispricings as they appear.
• A real foundation for AI: Liquid Grape's agents sit on a clean market surface, which is what turned experimentation into a workflow Henrik calls life changing.

Products used in this case study

What's Next

Liquid Grape is building its own MCP layer on top of Wine Labs, bringing internal systems, private offers and the wider fine wine market under one interface for agents, tools and customers. We are iterating with the team on richer historical coverage and the primitives that make AI-native wine operations work.