memtrust leaderboard

Standardized, reproducible benchmarks for agent-memory backends.

Backend	LongMemEval	LoCoMo	Contradiction (flagged / overwrite / stale)	Last run
Loading data.json...

This page reads data.json in this same directory, which is exactly the shape memtrust run produces. No build step, no framework -- open this file over any static file server (browsers block fetch() against file:// URLs, so use e.g. python -m http.server from this directory rather than double-clicking the file) and it renders whatever is in data.json, including a clearly-marked example when no live run has been published yet.