How to make a capsule

Capsules hosted on Galvoro are self-contained HTML mini-apps designed for higher education. You’re free to make and share any capsule you want, provided it meets a few requirements:

Capsules must teach or demonstrate a concept in higher education. Pointless games or generic puzzles don’t fit the platform.
Capsules must be short and focused. Galvoro is not the place for lecture slides, textbooks, or hour-long lessons — it’s the place for the interactive components that fit between or alongside them.
Capsules cannot save data between visits — each session starts fresh. Capsules may use approved third-party libraries and fonts (loaded from the CDN allowlist enforced by the platform CSP) but cannot call external APIs, services, or analytics.

What kind of capsule are you making?

While you can upload any mini-app relevant to higher education, Galvoro focuses on five types:

PLAY/LEARN— Pick-up-and-play games with an attached LEARN mode that explains the concept behind the game.

EXPLAIN— Animated mini-explainers, typically focused on a single concept and not interactive.

EXPLORE— Explorable explanations: longer, interactive learning sessions where the interaction is the lesson.

STUDY— Tools for studying or self-testing. Gamified flashcards, exam-prep quizzes.

TEACH— Built for in-class use — lecturer-driven games and polls on the lecture screen.

Uploaded capsules are categorized as MISC until a curator confirms they fit one of these types.

Building a capsule

You can write a capsule yourself, but AI coding assistants make this significantly easier. We currently recommend:

Claude Code — terminal-based AI coding assistant from Anthropic

ChatGPT Codex — OpenAI’s coding tool

GitHub Copilot — paid via GitHub’s Teacher/Student pack

Cursor — AI-first code editor

Local models via Ollama — we recommend the latest Qwen3-Coder or DeepSeek Coder model, run alongside an editor like VS Code

Working with an AI coding agent

Context matters. Before starting:

1.Download the Capsule Specification:

Download capsule-spec.md

2.Place it in an empty folder. Add any other context your AI tool needs: lecture notes, example datasets, references.
3.Open that folder in your AI coding tool.

A good starting prompt:

You are helping me build a capsule — a self-contained HTML mini-app — for the Galvoro platform.

Read `capsule-spec.md` in this folder first. It defines the rules, required files, allowed resources, and `capsule.json` schema your output must conform to.

  - Type: [TYPE]
  - Concept: [CONCEPT]
  - Audience: [UNDERGRAD/GRAD] students in [FIELD], course on [TOPIC]
  - Description: [DESCRIPTION]

Build a single self-contained folder with index.html at the root, all files relative, and a complete capsule.json manifest. No external APIs, no analytics, no persistent storage between sessions.

When you’re done

Create an account and head to the upload capsule page.

If your capsule is in a public GitHub repository, you can import it from GitHub instead of uploading a zip.

After you submit, automated checks run within a minute: file types, size, forbidden code patterns and external resources. A curator then plays your capsule in a sandbox before it is published. This typically takes a few days, and you’re emailed at each step.

A note on current limits

Galvoro doesn’t yet support multiplayer capsules, persistent storage between sessions, synchronous classroom modes (lecturer-and-student split apps), in-capsule analytics, or capsule-to-capsule communication. These are planned for later releases.