Cesium Developer Conference 2026 — Workshop

Building Your First AI-Driven Geospatial App with LLMs and Cesium

This repository contains all workshop materials for the Cesium Developer Conference 2026 hands-on session.

Quick start

All lab instructions, code, and assets live in the workshop/ folder.

👉 Go to the workshop README

Repository structure

workshop/
├── LAB_1.md                # Lab 1 instructions
├── LAB_2.md                # Lab 2 instructions
├── LAB_3.md                # Lab 3 instructions
├── LAB_4_challenge.md      # Lab 4 challenge instructions
├── README.md               # Workshop overview, setup, and schedule
├── TROUBLESHOOTING.md      # Common issues and fixes
├── lab1_lab2/              # Starter app for Labs 1 & 2
├── lab3_lab4/              # Starter app for Labs 3 & 4
├── lab4_mcps/              # Reference MCP server implementations
└── labs_completed/         # Completed reference solutions

Prerequisites

See workshop/README.md for full setup instructions.

How it works

sequenceDiagram
    actor User
    participant Browser as Browser<br/>(React + CesiumJS)
    participant Server as Server<br/>(Next.js /api/chat)
    participant LLM as LLM API<br/>(OpenAI / Anthropic)
    participant MCP as MCP Server<br/>(POI / Weather)

    User->>Browser: "Find cafes near the Colosseum"
    Browser->>Server: POST /api/chat<br/>{ messages, tools (JSON schemas) }
    Server->>LLM: streamText({ messages, tools, system })
    LLM-->>Server: stream: tool-call get_points_of_interest
    Server-->>Browser: stream: tool-call event
    Browser->>MCP: execute get_points_of_interest(lat, lon)
    MCP-->>Browser: [ { name, lat, lon }, … ]
    Browser->>Server: POST /api/chat<br/>{ messages + tool result }
    Server->>LLM: streamText (next step)
    LLM-->>Server: stream: tool-call flyTo + addEntity
    Server-->>Browser: stream: tool-call events
    Browser->>Browser: execute flyTo(lat, lon)<br/>execute addEntity(label)
    Browser->>Server: POST /api/chat<br/>{ messages + tool results }
    Server->>LLM: streamText (final step)
    LLM-->>Server: stream: text delta
    Server-->>Browser: stream: text delta
    Browser-->>User: Globe animates + text reply

AI inference runs server-side so API keys never reach the browser. Tool execute functions (camera, entities, etc.) run client-side against the live CesiumJS viewer.