Frequently asked questions
The specific stuff, not the generic stuff.
Safety & privacy
Does gitsparse ever see my code, or write or push anywhere?
No. gitsparse has no backend — every request goes straight from your browser to GitHub's own api.github.com and raw.githubusercontent.com. It only ever reads; it has no permission to push, write, or modify anything, even if you add a token. See the privacy policy for the full breakdown.
Do I need to sign in with a GitHub token?
No, browsing works without one — GitHub just caps unauthenticated requests at 60/hour, shared across everyone on your network. Adding a token on the sign-in page raises that to 5,000/hour, just for your browser session. The token is stored only in session storage, cleared the moment you close the tab, and never sent anywhere but GitHub's own API.
What does the cookie banner actually control?
Just one thing: whether Google Analytics loads. gitsparse itself sets no tracking cookies. Declining, or not choosing yet, means Analytics never loads at all — not just hidden, never requested. Full detail on the privacy policy.
What it can and can't do
Is there a size limit?
Yes — 10MB per download, shown as an estimate next to the Download ZIP button before you commit to it. If GitHub's own API response for a folder gets truncated (very large trees), gitsparse treats that as over the limit too, rather than risk shipping an incomplete zip.
Does it work on private repos?
Yes, with a token scoped to it. On the sign-in page, create a token with "Repository access" set to the private repo you want (not the default Public Repositories), then add it there — browsing and downloading both work exactly the same way after that.
Can I download a single file, or just folders?
Folders (and whole repos) only, for now — paste a .../tree/branch/path URL. Single-file .../blob/... links aren't supported yet.
What happens to the LICENSE file if it's outside the folder I asked for?
gitsparse includes it anyway. If GitHub detects a license at the repo root and your folder doesn't already have its own file by that name, the root license gets added to your zip too, so that context isn't lost. See Licenses & liability for the full policy.
vs. plain git
Why not just git clone?
You can! git's own sparse-checkout does this properly, and it's often faster for bigger folders. The terminal instructions page has the exact commands, and how it works covers what's actually happening under the hood. gitsparse exists because I could never remember those commands — if you can, the CLI is the better tool.
When does gitsparse make more sense than the CLI?
When you just want to look at or grab a small folder without a terminal handy — a phone, a locked-down machine, or you just don't want to think about git flags. For anything bigger, or if you need full history, the CLI route is usually better.