← Back to Work

Work / 03 of 04

AI & Agentic Tools

AI Building Blocks, MCP tools, generative passthrough and predictive input turn intelligent capabilities into useful applications.

These projects bring AI into working applications: vision, language and speech in Unity, MCP tools for conversational apps, generative passthrough and predictive typing.

Let context shape the experience

Use what an application can see and understand to shape how it responds. AI should serve the experience and make its behaviour meaningful.

Make capabilities easier to use

Connect complex models to applications through reusable code, clear interfaces and practical documentation. Make it easy to try an idea and understand the result.

Practical interfaces & tools

Predictive typing and MCP integrations connect capabilities to real tasks. The useful question is what the person can do with the result, and how clearly the application responds.

01 Swiss Transit Explorer

An MCP server and ChatGPT app that turns Swiss transit data into practical travel decisions.

Recorded Swiss Transit Explorer demo: route planning through MCP tools and ChatGPT widgets.

Our contribution

Starting on a train ride to Italy, we built an MCP server with mcp-use and Swiss public transport open data, then developed it into a route-planning app inside ChatGPT. We designed focused tools for connections, station boards, disruptions and delays, added route ranking and transfer-risk scoring, and built interactive route and departures widgets. Optional weather and GTFS line mapping made the results more useful. Beyond the first working demo, we refined tool selection, arrival-time parameters and token usage, simplified the interface, and resolved production failures involving widget assets, build ordering and ESM imports.

02 AI Building Blocks

Reusable starting points for bringing object detection, language and speech into Meta Quest experiences.

AI Building Blocks artwork showing a person interacting with a virtual llama in mixed reality
Meta AI Building Blocks project artwork.

Our contribution

We led the AI Building Blocks project at Meta Reality Labs, connecting Unity code samples, developer tooling and documentation. The work gave developers practical starting points for integrating object detection, language and speech into Quest experiences. Bringing those pieces together made the path from an AI capability to a working spatial application easier to follow.

03 Decart XR

Real-time AI transforms passthrough to match the atmosphere of a video.

Decart XR — project documentation.

Our contribution

We contributed fixes and improvements to the open-source Decart XR samples. Building on its real-time diffusion pipeline, we developed an immersive video-player prototype that transformed passthrough to match a video’s atmosphere. The result explored how generative visuals could extend what was playing on screen into the surrounding room.

04 XR Keyboard with AI Prediction

Exploring predictive typing for spatial interfaces with Fleksy.

XR Keyboard with AI Prediction — project documentation.

Our contribution

We independently released a Unity XR keyboard, then collaborated with Fleksy to prototype its predictive typing SDK in XR. The work connected spatial key input with word prediction, exploring how suggestions could support text entry inside an immersive application. It brought an existing typing technology into a different interaction setting through a working prototype.

Explore next / 04

Developer Relations

Have a project in mind? Let’s talk ↗