I wanted to earn a first open-source bounty, so I pulled every open GitHub issue I could find with a bounty label or an Algora bot comment and looked at what was actually there.
What I found
Out of 430 open issues:
- 290 came from just 13 repos that mass-post "bounties" (one had hundreds of open bounty issues and was a month old).
- Most of the rest sat on repos with 0-2 stars, or repos created in the last two months, often promising USDC.
- 3 had already been paid out but were never closed.
- Only 22 were on established projects. 94 more I could not check before hitting GitHub's API rate limit.
And the real ones are crowded. Some examples from the 22:
- A $50 bounty with 52 people attempting it and 13 pull requests already claiming it.
- A $5 bounty with 14 attempts and 4 claims.
- A $25 feature request with 12 attempts since June and nothing merged.
- At the other end, a $2,340 bounty (WearOS support for microG) with only 2 attempts, because it is genuinely hard.
My takeaway: the label search is nearly useless on its own. Check the repo's age and stars, then count the /attempt and /claim comments before writing a line of code.
The list
I put the filtered list online with a 0-100 score for each bounty and a plain note on how crowded it is:
https://real-bounty-filter.vercel.app
The top 5 are free. The full list of 22, the rejected repos with the reason for each, and a CSV are CA$3.
Honest caveats: it is a manual snapshot from 20 November 2026 rather than live data, and bounty amounts are parsed from issue text so verify on the issue before you start.
If you hunt bounties, I would like to know what signals you use to decide one is worth your time.

