Concepts
When a user runs into something broken, Mushi Mushi records it, explains it in plain English and can hand your editor a fix, without another dashboard to babysit.The loop in plain terms:
- Capture: users report what felt broken (by shaking the phone, tapping, or using the widget), and Mushi Mushi adds a screenshot and what they were doing.
- Plain read: how serious it is, what kind of problem it is and the likely cause, in everyday English.
- Connect: similar reports merge into one item, and the parts of the app that break most often stand out.
- Fix: optionally, an AI agent drafts the fix as a pull request (a proposed code change); you merge it, edit it or ignore it.
- Verify: an automated replay checks the fix before you ship it.
- Remember: past fixes become rules your editor sees on the next pull request.
Primary paths
- Architecture: the parts of Mushi Mushi and how data flows between them
- How classification works: how a report is filtered, then sorted and explained, in two passes
- How fixes get drafted: how an AI agent writes a fix and opens a draft pull request
Everything else
Knowledge graph · Inventory & gates · Orchestrator hooks · Quality scoring · Anti-gaming & reputation · Trigger modes · Rewards · Closed-loop evolution · Credentials · Runtime config
Last updated on