AI Search for GitHub

Search across PRs and issues. Onyx indexes your GitHub organization and answers engineering questions with cited sources.

Why did we switch from REST to GraphQL for the billing API?

The migration was decided in PR #847. The main reasons were reducing over-fetching on the invoice page and enabling the mobile team to query exactly what they need.

PR #847 — Billing API migration

Your engineering knowledge is scattered across GitHub

Architecture decisions live in PR descriptions. Bug context is in issue threads. GitHub's search is powerful for code but weak for knowledge.

• PR descriptions contain critical context that's impossible to find later.
• Issue threads have debugging solutions that new engineers can't discover.
• Engineering context is spread across hundreds of PRs and issues.
• GitHub's search finds code matches, not answers to engineering questions.

How teams use Onyx with GitHub

Find prior art for engineering decisions

Search PRs and issues to find how a similar problem was solved before, what the tradeoffs were, and who worked on it.

Onboard engineers faster

New engineers ask Onyx about the codebase and get answers sourced from PRs and issue threads.

Debug with historical context

Search issue threads and PR discussions to find how a similar bug was diagnosed and fixed in the past.

Cross-repo discovery

Find relevant PRs and issues across your entire GitHub organization, not just one repo.

Works great with

Onyx searches across all your connected tools. Combine GitHub with these connectors for unified AI search.