Homestead

A mastery-based homeschool platform for ages 5–13.

Homestead turns an open, connected curriculum of ~1,590 micro-topics into a clear learning path where every idea is genuinely mastered before the next begins — and the whole experience adapts to the child. It's built for a homeschooling parent teaching one or more children.

Live at homestead.puter.site.

Highlights

• Mastery ladder — Topics → Sections → Subjects, each gated at 90%+ so learning always builds on solid foundations.
• Connected timeline — every topic shows its prerequisites and what it unlocks, straight from the curriculum's dependency graph.
• Ready-to-teach lessons — full plans with parent notes, plus low-prep, printable “print & go” materials (worksheets, flashcards, and more).
• Verified tests — topic, section, and subject mastery tests, written and then independently re-solved so answers are trustworthy; digital or printable.
• Active recall — spaced-repetition memory cards per topic, with due reviews.
• Spaced practice — mastery-test questions a child misses are queued and resurfaced on an expanding review schedule until they stick, extending spaced repetition from facts to problem-solving.
• Adaptivity — timed challenges and parent-approved difficulty increases when a child excels.
• Adaptive daily calendar — a day-by-day plan you can reschedule, mark done, and add extra practice to.
• Records & voice analysis — log observations and record lesson conversations with live transcripts and AI coaching for the parent.
• Homestead Helper — an AI chat coach for parents, grounded in the child's real progress.
• Gamified for kids — XP, levels, collectible badges, celebration effects, and a colorful full-screen Kid Mode.
• Commune (shared teaching) — team up with other families in a private “commune,” approve and share what a child is working on for a given day, and cover each other's kids with a one-tap printable Day Sheet. Only the day's topics and an optional note are shared; each family's data stays in their own account.
• Insights & notifications, a streak tracker, and a downloadable guide.

Research-Backed Evidence

Homestead is built on learning-science foundations, not ad hoc design choices. Each one below comes with its main caveat too — the lab effect and the classroom effect aren't always the same size.

• Mastery learning — requiring students to reach a high level of proficiency (Homestead uses 90%+) on one unit before advancing to the next produces large, consistent achievement gains, especially for struggling learners. See Kulik, Kulik & Bangert-Drowns (1990), Effectiveness of Mastery Learning Programs: A Meta-Analysis, Review of Educational Research — SAGE Journals. Caveat: the largest effect sizes come mostly from shorter studies using experimenter-made tests aligned to the intervention; effects measured with independent, standardized tests over full courses tend to be smaller (though still positive). Mastery gating also only helps if there's a real remediation loop behind it, not just a repeated pass/fail gate.
• Active recall (retrieval practice) — testing yourself on material, rather than re-reading or re-watching it, produces stronger and more durable learning than nearly every other studied technique. See Dunlosky et al. (2013), Improving Students' Learning With Effective Learning Techniques, Psychological Science in the Public Interest — APS summary. Caveat: the effect is strongest for material tested the same way it was practiced (facts, definitions, recall-type questions); it transfers less reliably to novel problem-solving or far-transfer tasks, which is part of why Homestead also spaces problem-solving retries, not just flashcards.
• Spaced (distributed) practice — spreading study of the same material over expanding intervals, instead of massing it into one session, is the other top-rated technique alongside testing. See Cepeda, Pashler, Vul, Wixted & Rohrer (2006), Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis, Psychological Bulletin 132(3): 354–380 — author's copy — and Hattie & Donoghue (2021), A Meta-Analysis of Ten Learning Techniques, Frontiers in Education — open access. Homestead's memory cards use a Leitner-style expanding schedule, and missed mastery-test questions are now queued for spaced retry the same way, extending the effect from facts to problem-solving. Caveat: the effect is less bulletproof in real classrooms than in the lab. A 2025 study in real primary-school settings found the retrieval (testing) effect held up, but the spacing-interval manipulation itself did not reach significance for those students — see Retrieval practice enhances learning in real primary school settings, whether distributed or not.

Commune — shared teaching

Homeschooling parents often share teaching duties. Commune lets a small group of families do that inside Homestead without giving up their privacy.

How to use it (in the app):

1. Open the Commune tab and either create a commune or join one with an invite code another family shares.
2. Share a child's day. For a day another parent is covering your child, choose “Share today's focus,” pick the child, the date, and the topics they should work on, add an optional note, and share it with the commune.
3. Cover and teach. When you're covering, the “Covering today” section shows each child shared with you. Choose Teach on a topic for the full lesson, or Print day sheet for one packet covering every child you have that day — lessons, activities, resources, and the sharing parent's note.

What stays private: only the topics a parent explicitly approves for that day (plus an optional note) are ever shared. Progress, mastery, records, and recordings never leave a family's own account. The covering parent's app rebuilds each lesson locally from the shared topic, so what and how to teach is available without ever exposing how the child is actually doing.

Commune runs on a Puter serverless worker that stores only commune membership and the opt-in daily cards — never any family's learning data.

Tech

• Static front end: HTML + CSS + vanilla JavaScript (ES modules), styled with Tailwind (CDN).
• Backend, auth, storage, and AI via Puter.js.
• No build step. The site is served from the src/ directory — src/index.html is the entry point.

Running locally

Any static file server works, pointed at src/, for example:

npx serve src
# or
cd src && python3 -m http.server 8000

Then open the served URL. Sign in with a Puter account when prompted; each parent's students, progress, records, and recordings are stored privately in their own account.

Deployment

Pushes to main deploy automatically to homestead.puter.site via the Deploy to Puter GitHub Actions workflow, which runs scripts/deploy-puter.mjs.

The script uploads src/ to Puter hosting with its directory structure preserved, into a fresh release-<timestamp> folder, then atomically re-points the homestead subdomain at it (zero downtime) and prunes old releases. It uses the @heyputer/puter.js SDK directly rather than the Puter CLI, because the CLI flattens nested directories (so src/js/views/dashboard.js would 404). You can also run it from the Actions tab via workflow_dispatch.

Deployment requires a PUTER_AUTH_TOKEN repository secret (generate one at puter.com/dashboard → Create token). If it's missing, the workflow fails fast before touching the live site.

The curriculum data

The curriculum is not bundled — Homestead fetches the latest data at runtime from the open-source Marble Skill Taxonomy:

• https://github.com/withmarbleapp/os-taxonomy

This means the curriculum stays up to date automatically, and the app notifies you when topics are added or removed upstream.

Licensing

• Application code: MIT — see LICENSE.
• Curriculum data: licensed separately by its authors (ODbL 1.0 for the database, CC BY-SA 4.0 for the text) and must be attributed. See NOTICE and DATA-LICENSE.md for details and obligations.

Using the taxonomy inside this app is a "produced work," so the MIT license on the code is compatible — you only owe attribution (and share-alike if you ever redistribute a modified copy of the dataset itself).

Required attribution

Marble Skill Taxonomy (v1) · © Generative Spark, Inc. (Marble) · https://withmarble.com · licensed under ODbL 1.0 (database) and CC BY-SA 4.0 (content).

Contributing

Issues and pull requests are welcome. By contributing, you agree your contributions are licensed under the MIT License. Please don't commit copies of the taxonomy dataset into this repository — Homestead loads it at runtime.