A small-team AI tool for sorting Linear tickets
The author says their small team has too many Linear tickets, and new developers struggle to find the right context. They tried several AI tools, but most did not meet their needs, while Devin worked best but was expensive and did too much. They built a tool over a few weeks to triage tickets, find duplicates, check related PRs and GitHub issues, send Slack updates, track LLM traces, and watch cost.
Key points
- The team uses Linear and has many tickets across projects, services, repositories, and resources.
- The author says Devin was the best tool they tried, but it was expensive and did more than needed.
- The planned flow uses a cheaper model for triage and a heavier model for planning.
- The tool checks for duplicate tickets, existing PRs, GitHub issues, and the right repo for each team.
- It includes Slack updates, human review, LLM traces, cost tracking, and support for open source models.
Quick term guide
- developers
- Developers are people who build software, apps, or websites.
- GitHub issues
- Items used on GitHub to track bugs, tasks, or requests.
- AI agent
- An AI program that can inspect information and suggest what to do next.
- workflow
- A repeatable set of steps for getting a task done.
- repositories
- Places where a project’s code and related files are stored.
- resources
- The usage amount or compute limit spent while running an AI tool.
- open source models
- AI models whose code or model details are made available for public use or adaptation.
- open source
- Software whose code is available for people to view and often modify.