typeful-triage
Multiplayer GitHub triage dashboard where Jev answers kind, severity, urgency, duplicate and next step per issue.
Overview
typeful-triage is a shared dashboard for maintainers of public GitHub repositories. Issues and pull requests sync into Postgres and are replicated live to every browser with Rocicorp Zero. A TypeSafe System One model answers a fixed set of typed questions about each issue: kind, severity, urgency, whether it duplicates another, and above all what a maintainer should do next.
People confirm or correct those answers, claim issues and mark them done, and every correction is kept and shown to the model on later runs. Views include a personal work queue, triage grouped by next step (ask the author, reply, investigate, needs a decision, accept, close, waiting), an Unsure list ordered by lowest confidence, pull requests ordered by attention, repository health and an operator view with token spend and calibration against human feedback. Nothing is written back to GitHub, and a copy-prompt button hands items to Claude Code or Codex.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | What kind of issue is this? | — |
| Score | How severe and how urgent is this issue? | — |
| Yes / no | Does this issue duplicate another? | yes · no |
| Choice | What should a maintainer do next? | ask the author · reply with an answer · investigate · needs a decision · accept into backlog · close · waiting |
How it works
- Sync issues and PRs from GitHub into Postgres.
- Jev answers typed triage questions per issue, with past corrections as context.
- Maintainers review queues, confirm or correct answers and claim work.
- Low-confidence or disputed items appear in the Unsure view.
Strengths
- Human corrections are kept and fed back to later runs.
- Unsure view surfaces low-confidence decisions first.
- Read-only toward GitHub; it is not a bot.
- Tracks token spend and calibration against human feedback.
Caveats
- Requires Postgres, Zero and a TypeSafe key to run.
- Designed for public repositories.
At a glance
- Type
- App
- Status
- beta
- Runtime
- web app with Postgres and Rocicorp Zero
- Has code
- yes
Use cases and integrations
How curated lists describe it
Open-source maintenance: multiplayer triage dashboard where Jev answers a fixed set of typed questions per issue — kind, severity, urgency, duplicate, and next step — and every human correction is kept and shown back to the model on later runs.
yibie/awesome-jevUses typeful jev, zero sync to pull and sync large repositories for issue triage.
hellogumbo/awesome-jev
Frequently asked questions
What does Jev decide?
Each issue's kind, severity, urgency, duplicate status and recommended next step.
Does it change anything on GitHub?
No, it is a shared view; replies and merges still happen on GitHub.
How does it improve?
Human corrections are stored and shown to the model on later runs.
Related projects
CLI, MCP server and skill for issue triage where Jev predicts issue kind, affected module and whether information is sufficient.
MCP serverprototypePythonGitHub Action that triages issues with Jev, applying label, spam, needs-info and duplicate tags only above set thresholds.
ProjectbetaTypeScriptOne YAML task to classify, route, score or verify text with rules, Jev, NanoJev or an OpenAI-compatible model, with cost records.
CLIbetaPythonPython sandbox turning Jev issue-category and reproduction judgments into testable routing proposals, with offline replay.
ProjectprototypePythonPortuguese-language Go proof of concept that tests whether Jev's confidence is calibrated enough to auto-route support tickets.
DemoprototypeGoOpen-source team agent platform for Slack, GitHub and more that asks Jev once per session which agent runtime and model to use.
ServicebetaTypeScript