Qloo: The Cultural AI Behind Taste-Aware Agents
Qloo is a New York company that models taste: how a liking for one film, artist, restaurant or brand connects to likings in completely different categories. Its API takes whatever a product knows about a person or audience, such as a favorite movie, a musician they follow or the neighborhood they’re in, and returns ranked predictions of what else they’ll enjoy, from restaurants and hotels to podcasts and brands. Qloo has spent more than a decade building this, and it now pitches itself as “the privacy-first cultural intelligence layer for AI,” feeding taste into language models and agents without ever needing a name, email or device ID.
The quick answer: Qloo is one of the oldest and best-funded companies in taste data, with customers like Netflix, American Express, Live Nation, Michelin and Match Group. It’s built for businesses that want recommendations and audience insight across categories, and increasingly for AI agents that need to choose well for a person. Unlike memory layers, it deliberately doesn’t keep profiles of individuals. A disclosure: we make Carly, an AI assistant with its own memory, so we overlap on the personal-context side. Everything below comes from Qloo’s site, documentation, announcements and press coverage.
Qloo at a glance
Checked against Qloo’s site, developer documentation and press releases on November 12, 2026.
| Qloo | |
|---|---|
| What it is | A taste and cultural intelligence engine, sold as an API |
| Coverage | More than 600 million modeled entities: places, brands, titles, artists and more |
| Categories | Music, film, TV, dining, travel, fashion, brands, podcasts, books, sports and places |
| Main API | Insights API, ranked by affinity scores from 0 to 100 |
| For AI tools | Qloo MCP Server (beta) and a Documentation MCP Server |
| Privacy | No personally identifiable information used or required |
| Company | Founded 2012 in New York by Alex Elias (CEO) and Jay Alger (COO) |
What Qloo does
Qloo’s homepage describes its foundation as “the world’s most detailed cultural graph” paired with “proprietary taste intelligence.” In plain terms, it knows a great deal about things, such as what a restaurant, album or hotel is like, and about how people’s affinities connect them. Someone who likes a particular indie film director may be likely to enjoy a certain style of restaurant, and Qloo is built to find those links across categories.
It sells six capabilities built on that foundation, as its capabilities page lays out:
- Enrichment. Matching a company’s own records, like a list of venues, to Qloo’s entities and attaching their attributes and relationships.
- Understanding. Explaining what defines a combination of interests, place or brand, and what other affinities it suggests.
- Recommendations. Ranking what fits across dining, travel, entertainment, brands and places, starting from even a single interest.
- Locality. Reading the taste of a place at any scale, down to a few hundred feet.
- Grounding for AI. Giving language models and agents real-world facts and a person’s taste in one call.
- On-device models. Compact models that rank and match on the phone, without sending context anywhere.
Taste for AI agents
Qloo’s newest pitch is to the companies building AI assistants. Its argument is that an agent acting for someone needs to understand the person, not just the instructions. “Qloo is a partner to the model API, not a chat product,” its capabilities page says. The model handles the conversation, and Qloo handles the facts and the taste behind a choice, as a tool the model calls once.
Qloo’s illustration is an agent booking dinner for one night in Copenhagen. Instead of crawling restaurant websites one page at a time, the agent asks Qloo once, gets venues with real hours, locations and attributes, and has them ranked by what the person’s interests point to, even when those interests have nothing to do with food.
For developers, Qloo offers two MCP servers that connect tools like Claude, Cursor and Windsurf: a beta server for querying Qloo’s data directly, with prompts like “Recommend five restaurants in Brooklyn for someone who likes natural wine and A24 films,” and a documentation server that helps an AI tool write Qloo API requests. Qloo also ran an agentic hackathon in 2026 under the banner “Agents, but with taste.”
How the API works
The core product is the Insights API. You send what you know, such as an entity (a movie, an artist), tags (a genre, a cuisine), a location, or audience signals like age group, and ask for results in a category. Qloo returns a ranked list with affinity scores, where 100 means a strong relationship and 0 means none. Other endpoints cover search, trending entities, audiences and comparisons between groups.
Access starts with an API key from Qloo’s team, passed in a request header. Qloo doesn’t publish pricing. In 2024, Elias told TechCrunch that the company charges a monthly subscription for API access and that contracts start in the five figures.
Privacy by design
Qloo’s privacy stance is central to how it sells. Its homepage says Qloo “was built from the beginning around a simple premise: cultural relevance can be inferred from entities, relationships, and context without maintaining identity-level consumer profiles.” It says there’s no consumer identity graph, no names, emails, phone numbers or device IDs required, and each request is judged from the context supplied at that moment.
As for where the data comes from, Qloo’s FAQ lists its own cultural ontology, first-party data from platforms such as TasteDive (a recommendation site Qloo bought in 2019), direct partnerships, licensed metadata, open data, and learnings from API usage where agreements allow. That’s more detail than Elias gave TechCrunch in 2024, when the reporter noted he wouldn’t say where the data came from. Qloo refreshes data daily, weekly or monthly depending on the type.
Who uses Qloo
Qloo’s September 2026 announcement names American Express, Live Nation, Match Group, Michelin, Netflix and JCDecaux among its customers. Some published examples:
- Match Group. Tinder and Match’s Affinity Apps (BLK, Chispa, Upward and Yuzu) use Qloo to let people build profiles around what they’re into. According to Qloo’s case study, Yuzu users who added the Qloo-powered “Currently Into” feature saw a 70% increase in Likes.
- Michelin uses Qloo for personalized recommendations in its travel and dining guide.
- Starbucks, Hershey’s and Netflix, per TechCrunch’s 2024 coverage, used Qloo for neighborhood-specific in-store playlists, customized candy assortments and identifying actors who resonate with particular audiences.
- Kearney, the consulting firm, used Qloo’s models to study consumer tastes, including links between electric-vehicle buyers and off-road enthusiasts.
These are company-reported results. Qloo sells to businesses, and we found no independent customer reviews to weigh against them.
The company
Alex Elias started Qloo in 2012. While studying law at NYU, he had noticed that taste knowledge was split across companies like Spotify and Netflix with no system connecting it across domains. He co-founded it with Jay Alger, who previously led the digital agency Deepend and is now Qloo’s chief operating officer. Qloo began with a consumer app before turning to enterprise customers, with Twitter among its first in 2015.
Qloo raised a $25 million Series C in February 2024 led by AI Ventures, with AXA Venture Partners, Eldridge and Moderne Ventures. In July 2024, Bluestone Equity Partners invested $20 million, which AlleyWatch reported brought Qloo’s total funding to $77 million. In September 2026, Qloo said a Gartner report on synthetic population and behavioral simulation placed it as one of two front runners among the 33 startups profiled.
How Qloo compares
Qloo sits next to memory and personalization tools for agents, but it works differently. Memory layers like Supermemory, Honcho, Zep and Cognee store what they learn about each user over time. Qloo stores nothing about the individual and brings the taste knowledge instead: give it a few signals and it predicts the rest from its model of culture. TasteGraph is the closest newer rival, focused on turning a person’s taste into shopping choices, and Rollup learns interests from a person’s camera roll on their phone.
The two approaches can work together. If you’d rather have an assistant that actually remembers you, Carly has memory built in: it saves named memory documents that it reads and searches in later conversations, searches a knowledge base of files you upload by meaning, and keeps your to-dos, data tables and CRM in one workspace. Because it already knows your preferences, there’s nothing to re-explain, whether you use it inside ChatGPT or Claude or by text, and it also labels and sorts email across all your accounts, handles advanced scheduling, and makes and answers calls and texts.
Who Qloo suits
Qloo is a good fit if:
- You run a consumer business in travel, dining, entertainment, media, retail or financial rewards and want recommendations that cross categories.
- You’re building an AI agent that recommends places, products or media and want it grounded in real taste data.
- You need personalization without collecting or sharing personal identifiers.
- You want audience and location insight for marketing, merchandising or site selection.
Look elsewhere if:
- You want memory of what each individual user said or did; Qloo doesn’t keep that.
- You need self-serve sign-up and published pricing.
- You’re a solo developer looking for a free tier.
FAQ
What is Qloo?
Qloo is a cultural AI company that models taste across music, film, dining, travel, brands and more. Its API predicts what people are likely to enjoy from signals like a favorite movie or a location, without using personal identifiers.
What is Qloo’s Taste AI?
Taste AI is Qloo’s name for its taste intelligence engine: a knowledge graph of more than 600 million entities combined with models of how people’s affinities connect them across categories.
Does Qloo use personal data?
Qloo says it doesn’t use or require personally identifiable information, keeps no consumer identity graph, and evaluates each request from the context it’s given.
Does Qloo have an MCP server?
Yes. Qloo offers a beta MCP server for querying its data from tools like Claude and Cursor, and a documentation MCP server for building with its API.
Who founded Qloo?
Alex Elias and Jay Alger founded Qloo in New York in 2012. Elias is CEO and Alger is chief operating officer.
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
