The README first-fold restructure is being handled separately, so drop it
from this PR rather than block the two changes that do not depend on it.
README.md and mem0-ts/README.md go back to main verbatim. What stays:
- docs/migration/oss-v2-to-v3.mdx and platform-v2-to-v3.mdx still said
91.6 LoCoMo / 93.4 LongMemEval. main's README already publishes 92.5
and 94.4, so the migration guides were the stale side of that
disagreement, not the README.
- the product one-liner is unified across the 11 other places that each
worded it differently.
README first fold:
- banner 800px -> 520px
- badges condensed from two paragraphs into one row (Trendshift kept, inline)
- lead with Introduction instead of benchmarks
- new How it works section explaining the add/search loop
- benchmarks moved below the intro and retitled from the date-stamped
New Memory Algorithm (April 2026)
- dropped Research Highlights, which restated the benchmark table verbatim
Benchmarks: the README was already correct at 92.5 / 94.4, matching
mem0.ai/research and docs/core-concepts/memory-evaluation.mdx. The stale
copies were in the migration guides, which quote the numbers as a live
reason to upgrade. Updated both to 92.5 / 94.4 (+21 / +27).
docs/changelog/highlights.mdx keeps 91.6 / 93.4 inside its dated
2026-04-14 entry, which records what was announced at the time.
One-liner: standardized on 'the memory layer for AI agents', already the
canonical form in cli-spec.json, cli/python/pyproject.toml and the CLI
specification. Swept the remaining variants.
kartik-mem0