All projects

Book Mash

A multi-judge measurement engine for book manuscripts — six independent judges score prose on craft and epistemic dimensions, producing a JSON ledger and report.

CompanyIndependent
Timeframe2026 - Present
FocusDeveloper Tools & AI Agents
Developer Tools & AI Agents

Behind the build.

  • First consumer of the tool is From Copilot to Colleague, whose chapters are graded by a panel of rival open models.
  • Outputs a structured, machine-readable scoring ledger rather than a single opaque rating.
AI EvaluationPublishing Tools

The next idea is waiting.

Back to the collection.