I was building an AI agent to manage my wine collection. (First world problem?😅) The agent needed structured wine data it could query to make suggestions based on what I had in storage and what I had previously enjoyed. I went looking for a wine data API. What I found was both inaccessible and priced for a different customer entirely, and none of it was built with AI agents in mind.
It's important to note that, that's not the same as saying nothing exists. There are, of course, several wine data services. The problem is what they are, who they're built for, and what they cost.
Wine-Searcher has an API. It's built for merchants, not developers. The Price Check tier starts at $250/month for 500 calls per day and goes to thousands per month for 5,000 calls per day. It returns price data and an aggregated critic score for a single, precisely named wine. It doesn't do search. It doesn't do food pairings. It doesn't return full tasting profiles or structured wine attributes beyond the basics. And it explicitly excludes PRO data — historical prices, for instance, are locked behind a separate subscription tier. If you want that for your application, you're paying for two subscriptions.
Liv-ex offers LWIN-based APIs for the fine wine trade. They require a Gold membership, and critic scores and notes require separate licensing agreements with individual publications. The target customer is a wine merchant or auction house, not someone building an app.
Community datasets — there are several publicly available wine datasets with reasonable coverage. But they're flat files with no API layer, no search, no enrichment, and no SLA. If you want to use them, you're building the infrastructure yourself.
Scraping is the other option, and it's not really an option. Rate limits, legal exposure, maintenance overhead, and inconsistent data structure make it a poor foundation for anything serious.
So the choice, for a developer, has been: pay enterprise prices for narrow functionality, or build your own infrastructure on top of static datasets. WineAPI is a third option: a developer-facing REST API with clean documentation, consistent schemas, reasonable pricing, and data that covers the full wine record — identity, classification, scores, prices, pairings — in one place.
We launched recently. The catalog is live and the enrichment pipeline runs continuously. Wine-Searcher and Liv-ex have years of data accumulation behind them, and we don't claim otherwise. What we offer that they don't is a product actually built for developers, at prices that don't require a procurement process — and one of the first wine data APIs designed to work natively with AI agents.
The Data
The catalog combines multiple data sources into unified wine records. A single record carries industry-standard identifiers, production attributes, grape variety data, regional classification, critic scores, retail prices, and food pairings in one place.
The catalog is not static. A background pipeline runs continuously, identifies records that need updating, and refreshes them automatically. No manual intervention.
Search
Standard keyword search fails immediately for wine queries. "Something earthy and medium-bodied from Burgundy" returns nothing useful if the system is just pattern-matching on text. The queries carry meaning that LIKE operators cannot reach.
WineAPI implements hybrid search — keyword matching and vector similarity running in parallel, results merged by confidence score. Every wine in the catalog has a semantic embedding. Queries get embedded at runtime and matched against the catalog by cosine similarity via a pgvector HNSW index. Natural language queries work. Exact name matches still rank first. If the embedding service goes down, keyword search keeps functioning.
Endpoints
Beyond search, the API provides:
- Wine detail — full record including production attributes, regional data, grape varieties, tasting notes, and scores
- Prices — current retail prices across merchants
- Price history — time-series price data across merchants (Pro tier)
- Pairings — food pairing recommendations per wine
- Identification — submit a label photo, get back a matched wine record; the pipeline runs OCR, searches the catalog, and disambiguates between candidates when confidence is close
- Reference data — regions and grape varieties for building filters and faceted navigation
AI Agent and MCP Support
WineAPI started as a backend for an AI agent, and that use case is a first-class citizen of the product. WineAPI ships with a Model Context Protocol (MCP) server, which means it can be connected directly to LLM-based agents and tools — including Claude, Cursor, and any other MCP-compatible host — without custom integration work.
This matters if you're building wine-aware AI features: a sommelier assistant, a cellar management agent, a restaurant pairing tool, a recommendation engine. The MCP server exposes the full API surface to your agent as native tools. The agent can search the catalog, pull wine detail, look up pairings, and query prices the same way a human developer would — through structured, documented calls rather than unstructured web access.
We believe WineAPI is one of the first wine data APIs built with this use case in mind from the start. The REST API and the MCP server share the same data, the same authentication model, and the same tier structure.
Using the API
Authentication is an X-API-Key header. Errors return a consistent JSON structure with a human-readable message, a machine-readable code, and a status code — the same format across every endpoint. The documentation is generated from the live OpenAPI spec, not maintained separately. The spec itself is public at /spec if you want to generate a client or review the contract before signing up.
Tiers are Free (100 req/day), Supporter (250 req/day), Pro (1,000 req/day, full feature access including price history), and Enterprise. No credit card required for the free tier.
Get Started
Docs: wineapi.io/docs
Spec: api.wineapi.io/spec
Enterprise and custom work: pinnacleridgesolutions.com
Pinnacle Ridge Solutions builds data infrastructure and API products. WineAPI is our first of many standalone products.