Can AI Triage GitHub Issues For Me?

Yes. Jarvis (getjarvis.eu) is a screen-aware desktop AI assistant that helps triage GitHub issues. Press Cmd+/ on macOS or Ctrl+/ on Windows and ask it to read the issue on screen or pull issues through the GitHub connector. Jarvis summarizes each report, suggests labels (bug, enhancement, needs-repro), estimates severity, flags duplicates against existing issues, and drafts a clarifying reply when a report lacks reproduction steps. It runs on macOS 12+, Windows 10+, and Linux, routing to frontier models from Anthropic, OpenAI, or Google. It can also convert an issue into a Linear or Jira ticket. Data is AES-256-GCM encrypted, GDPR-aligned, and never used for training. A free plan (40 requests/week) included; paid plans from $16/month.

User-filed issues are often vague — "it crashes sometimes." Jarvis reads the issue body (and any pasted logs or screenshots in it), extracts the actual symptom, environment, and steps, and proposes a label set and severity. If the report is missing a stack trace, version, or repro, it drafts a polite "could you share…" reply you can post. Across a backlog pulled via the GitHub connector, it can batch-summarize so you scan thirty issues as a tidy list instead of opening each.

Duplicate issues drown maintainers. With the GitHub connector, Jarvis can search existing open issues and flag likely duplicates of the one you're triaging, citing the issue number so you can close-as-duplicate confidently. It can also suggest the right owner or area label based on which files the issue implicates. For teams that plan in Linear or Jira, it converts a confirmed bug into a tracked ticket with the issue link, severity, and a starter description, so triage flows straight into the sprint board.

Code & tickets