SAAS PRODUCTS In production

Open-Source Issue Discovery Service

Developers who want to contribute to open source stall at discovery: GitHub search assumes you already know the repository, and good-first-issue labels are applied too inconsistently to filter on. We built a service that ingests issues hourly, filters by language and stack, and signs users in with the GitHub identity they already have.

100k+
issues indexed
Hourly
GitHub sync
100
daily active users

The problem we were asked to solve

Developers who want to contribute to open source usually stall at the same step: finding an issue that matches their language, their skill level and their available time. GitHub's own search is built for people who already know which repository they are looking at, and the good-first-issue label is applied inconsistently enough to be unreliable as a filter. The result is a long tail of maintainers waiting for help and a long tail of developers willing to give it, failing to meet because discovery is bad. Any service solving this has to stay current — an issue list that is a day stale sends contributors to work someone else already picked up.

What we built

01

Hourly GitHub ingestion

Fresh issues are pulled from the GitHub API every hour, keeping the index current enough that a contributor arriving at an issue finds it genuinely open rather than claimed hours earlier.

02

Language and stack filtering

Issues are filterable by programming language and stack, so a developer sees work they can actually start today instead of a general feed they have to triage themselves.

03

GitHub OAuth authentication

Sign-in happens through GitHub, matching the identity contributors already use and removing a separate account from the path between arriving and contributing.

04

Click-through tracking

Tracks which issues get taken up, which turns the service's own usage into signal about what kinds of issues actually convert browsers into contributors.

Web application GitHub API OAuth Hourly sync Analytics
Issue discovery screen: language, label and effort filters, a list of freshly synced GitHub issues with nobody assigned, and a pull-request helper walking from branch name to a drafted PR
Filtered issue feed with the pull-request helper alongside it. Reproduced with anonymised data.

What changed for the client

The service indexes more than one hundred thousand issues, refreshed hourly, with around a hundred developers using it daily to find work matched to their stack. The discovery step that previously ended most contribution attempts becomes a filtered list. It is live in production.

  • More than one hundred thousand issues are indexed with hourly refresh, so contributors reach issues that are still genuinely open.
  • Around a hundred developers use the service daily to find work matched to their language and stack.
  • GitHub OAuth sign-in removes a separate account from the path between arriving and contributing.
  • Click-through tracking turns the service's own usage into signal about which kinds of issues actually convert browsers into contributors.

Want similar results?

Tell us what the process looks like today and we will tell you what can be automated — and what should not be.

LET'S TALK