Bulk problem crawler
Pull every Codeforces problem statement into Supabase using your browser as the worker. cf-mirror can't scrape Codeforces from the server (Cloudflare blocks DC IPs), but your Firefox tab can. This page drives the userscript: it picks the next un-captured problem, opens it in this same tab, the helper captures it and redirects back. Repeat until done. No popups, no new tabs.
Progress
0%
This session
0
Status
paused
Consec. fails
0
Recent activity last 50 events
No activity yet. Hit Start to begin.
Settings
Each judge has its own queue. CF / CSES / AtCoder / LeetCode rows live in the same `problems` table.
seconds
Be polite to the judge — 2–5 s is sane.
How it works
- Click Start. This tab navigates to a CF problem page with
?cfm_crawl=1. - The companion userscript captures the statement & samples, POSTs them to
/api/capture, then sends this tab back to/crawl?cfm_crawl=1&done=<id>. - This page records success and after delay seconds, picks the next problem.
- Loops until the queue is empty, you Pause, or 5 fails in a row (likely CF rate-limit) auto-pause.
Tip: keep this Firefox tab focused. Background tabs run on a throttled timer in Firefox, which slows the loop way down. Set delay to 2 s, leave it overnight on a laptop with the screen off (lid open) and it'll do ~30 problems/min.