The Rebuild Index · retrieved September 26, 2026
Students keep rebuilding the same tools, faster than GitHub itself is growing.
GitHub holds 131 separate syllabus-to-calendar projects. From 2024 to 2025, new ones went from 3 to 43, a 14× jump, while new repositories across all of GitHub grew 1.4×. The same small tool, rebuilt from scratch, faster every year.
It is speeding up, not slowing down
For years a handful of these appeared annually. Then 2025 brought 43, and 2026 has already passed that with 3 months to go. One plausible reason, which the counts can't prove, is that AI coding tools turned a weekend project into an evening.
Seven kinds of student tool, measured against GitHub
Each row is a GitHub repository search. The last column divides each tool’s growth by GitHub’s own, so 1× means “growing exactly as fast as GitHub”. 6 of the seven are being rebuilt faster than GitHub is growing.
| Tool | Repositories | New in 2024 → 2025 | Growth | vs GitHub |
|---|---|---|---|---|
| All of GitHub (baseline) | n/a | 43,274,350 → 58,584,211 | 1.4× | 1× |
| Syllabus to calendar | 131 | 3 → 43 | 14× | 11× |
| University schedule planner | 134 | 10 → 27 | 2.7× | 2.0× |
| Citation generator | 593 | 66 → 144 | 2.2× | 1.6× |
| “For my university” (in the description) | 3,472 | 508 → 653 | 1.3× | 0.9× |
| Flashcard generator | 3,691 | 202 → 639 | 3.2× | 2.3× |
| GPA calculator | 7,693 | 943 → 1,674 | 1.8× | 1.3× |
| Student attendance tracker | 8,722 | 752 → 2,648 | 3.5× | 2.6× |
Built one of these yourself? Put it somewhere other students can use it. If publishing opens, the first 100 founding builders get their first tool reviewed and featured, as long as it meets the publishing rules.
Be a founding builderWhat this does and doesn’t show
- A repository is not a student. Some are tutorials, class assignments, templates or one person’s second attempt. This counts rebuilding, not people.
- Small numbers, big multiples. Syllabus-to-calendar went from 3 to 43 new repositories a year, so read its multiple as a direction, not a precise rate. The larger categories are steadier.
- Stars don’t set these apart. 98.5% of syllabus-to-calendar repositories have three stars or fewer, but so do 99.1% of all repositories GitHub created in 2025. Almost nothing on GitHub gets stars, so we don’t read anything into them.
- GitHub itself keeps growing, partly because AI coding tools make new repositories cheap, which is why every multiple here is compared with GitHub’s own.
- Search counts are GitHub’s own figures and can be approximate for large result sets. Forks are excluded, so a copy of someone else’s repository isn’t counted as a rebuild.
- The queries are keyword searches on repository names and descriptions, so they miss projects described differently and include a few that only mention the words.
Why we track it
educagents is building a place for exactly these tools: a student builds one for their own course, publishes it, and other students can use it instead of starting from zero, at a price the builder sets, including free. Publishing and pricing are designed, not built yet. This index is how we’ll know whether the rebuilding slows down once there is somewhere to put the first good one.
Method, and how to check it
Each figure is the total_count from GitHub’s repository search API for the query shown in the dataset, with created: date ranges for the by-year counts and stars:0..3 for the star split. The baseline is the same search with no keywords: every public repository created in each year. Retrieved September 26, 2026. We re-run it monthly and keep the date on every number.
The full dataset, including every query string, is at /rebuild-index/data.json, and the script that produces it is public at github.com/educagents/rebuild-index, so you can re-run every number yourself. You’re welcome to use the data with a link back (CC BY 4.0).
Built one of these yourself?
If publishing opens, early-access builders get it first. We email when there is something real to show, and not before.