Open Door

Open Door is a cultural onboarding agent that starts with what you already love and uses Qloo's cultural intelligence to help you discover where to go next.

Comment

Inspiration

A map can tell you where you are. It cannot tell you where to begin.

Moving somewhere new often means starting with recommendations that know the city but know nothing about you: the most popular attraction, the highest-rated restaurant, another list of “must-see” places. Those answers can be useful, but they rarely explain how an unfamiliar place connects to what you already love.

We imagined a different starting point: What if your favorite music, films, books, and everyday interests could help you find your way into a new place?

That became Open Door—a cultural discovery experience built around curiosity rather than popularity. Instead of asking people to define their entire personality, it starts with interests they explicitly choose to share and offers a few meaningful directions to explore.

The goal isn't to keep people inside the things they already know. It's to help them discover where their curiosity could take them next.

What it does

Open Door turns a few cultural interests into an interactive discovery journey.

Users provide their interests, a place, an intention, and how adventurous they want to be. Open Door presents up to three directions:

  • Familiar: a comfortable starting point.
  • Adjacent: a neighboring interest with room to explore.
  • Unexpected: a less obvious direction worth investigating.

These are Open Door's navigation labels—not Qloo scores or judgments about a person.

Each discovery includes an explanation and inspectable source information. Users can save a door, provide feedback, or open it as the starting point for another exploration.

The Discovery Lab

The Discovery Lab takes this further with six exploration strategies:

  1. Stay in Your World — explore closer connections to existing interests.
  2. Follow the Thread — follow an adjacent cultural interest into a new discovery.
  3. Cross Into Another World — explore possibilities across different cultural domains.
  4. Find Your People — explore experiences for a user-selected occasion or group.
  5. Take the Wild Door — pursue a less obvious cultural connection.
  6. Make It Real — shape discoveries into a practical exploration plan around user-provided constraints.

Users can save checkpoints, create independent branches, revisit previous discoveries, and compare different paths from the same starting point. Exploring something unexpected doesn't mean losing the direction they liked before.

A discovery can become a new starting point. A checkpoint can lead to several different journeys. Users can follow the path that interests them, return to an earlier decision, or try something completely different.

The core interaction is not “ask, answer, disappear.” It is “discover, inspect, choose, and continue.”

The experience also includes exploration-depth choices, editable anchor categories, context controls, a group example, saved doors, feedback, a shareable story, activity counts based on user actions, a door mascot, optional sound effects, and reduced-motion support.

How we built it

We built Open Door with React, TypeScript, Vite, Node.js, and Express, with a server-side integration path for Qloo Search and Insights.

The intended live discovery workflow is a bounded, structured orchestration:

  1. Interpret explicit intent. Use the interests, context, exploration depth, and feedback the user chooses to provide.
  2. Resolve cultural entities. Use Qloo Search to resolve typed interests into provider entities, or reuse an exact returned entity ID when continuing from a discovery.
  3. Retrieve connected possibilities. Make multiple Qloo Insights requests across relevant entity types.
  4. Evaluate and organize results. Deduplicate candidates, apply exclusions, and organize returned possibilities into navigable directions.
  5. Continue the journey. Use a returned entity as the next anchor so exploration can evolve rather than repeatedly starting from scratch.

Qloo is intended to provide the cultural retrieval layer, while Open Door provides the reasoning and orchestration needed to turn retrieved possibilities into an interactive journey. The current workflow does not use an external LLM to invent cultural relationships.

The Discovery Lab uses a shared discovery-provider interface, checkpoint snapshots, independent branch histories, and validated browser-local persistence. This allows the branching experience to operate in Demo Mode while live API credentials are pending.

We deliberately separated live and demonstration modes. A Live label requires a successful Qloo-backed journey. Failed requests remain visible failures or partial results; they are not silently replaced with sample recommendations.

Current integration status: The server-side Qloo integration path is implemented, but Qloo credentials have not yet been configured in this workspace. Successful live Qloo requests have not been verified here. Demo Mode uses authored sample journeys, not live Qloo results.

Challenges we ran into

Making cultural discovery agentic without making it opaque.

We wanted the workflow to choose and sequence useful retrieval actions while keeping those actions inspectable. Structured orchestration and visible source information help users understand how discoveries were produced.

Distinguishing discovery from unsupported promises.

A cultural relationship does not prove that a venue is open, affordable, nearby, accessible, or available. We keep user-provided planning conditions separate from verified facts and avoid presenting cultural entities as confirmed bookings or itineraries.

Preserving exploration without creating confusion.

Branching becomes unreliable if changing one path also changes its siblings. The Discovery Lab uses independent checkpoints and branch histories so users can explore alternatives without losing previous progress.

Keeping the experience welcoming without sacrificing usability.

The mascot and sound effects needed to support the product rather than distract from it. Animations are optimized, sound is optional and muted by default, and reduced-motion users receive still artwork.

Building responsibly while waiting for live credentials.

We wanted to develop and test the discovery experience without pretending that sample content came from Qloo. Separating the provider interface from the experience allows Demo Mode to remain useful while preserving a clear path toward real API integration.

Accomplishments that we're proud of

  • A distinctive cultural discovery experience: three initial directions lead into six exploration strategies.
  • A structured Qloo integration path: entity resolution, multiple Insights requests, and entity-based continuation are designed into the server-side workflow.
  • An interactive Discovery Lab: checkpoints, independent branches, comparison, and continuation let users explore alternate possibilities without losing their progress.
  • Clear evidence boundaries: authored examples, provider results, application decisions, and unverified planning assumptions remain distinguishable.
  • An approachable product experience: the door metaphor, mascot, responsive layouts, and accessible controls make exploration inviting.
  • A tested foundation: the project reports 38 state/server tests and 14 browser scenarios covering important workflows, including branch isolation, persistence, sound controls, mascot behavior, and mobile usability. These counts should be verified against the final submitted build.

Most importantly, we built an experience that gives people control over discovery. An unexpected suggestion is an invitation—not a verdict.

What we learned

We learned that cultural intelligence becomes most useful when people can act on it, question it, and change direction.

A large result set is not necessarily a better experience. A few understandable possibilities, a clear explanation, and a reversible next step can be more useful than another endless feed.

We also learned that identity and taste are not interchangeable. Someone choosing a musician, film, or book gives us an interest to explore—not permission to infer their background, sensitive characteristics, or cultural identity.

Finally, trust is part of the architecture. Keeping credentials server-side, preserving provenance, failing explicitly, and distinguishing examples from retrieved results are product decisions, not just implementation details.

What's next for Open Door

Our first audience is people beginning somewhere unfamiliar: recent movers, students, travelers, and anyone who wants to explore a place through interests they already care about.

Our next steps are to:

  • Evaluate discovery quality with real users: measure whether people understand the connections and find directions worth following, not just whether they click.
  • Benchmark cultural discovery: compare relevance, novelty, and explanation quality against generic recommendation approaches rather than assuming superiority.
  • Add verified local information: connect appropriate sources for opening hours, cost, accessibility, and availability before presenting actionable local plans.
  • Extend group exploration: move beyond the prepared group example using voluntarily provided interests and tested live retrieval.
  • Make saved journeys portable: support restoring exported Discovery Lab histories while preserving the account-free, browser-local experience.

Open Door's goal is not to tell people what they should like. It is to help them discover where their curiosity could take them next.

Start with what you love. Follow the connection. Open the next door.

Built With

Share this project:

Updates

Submission history