How the new leads read on GitHub.

The memory & context layer for AI agents: load only the context they actually need.

Your agents re-read their whole notebook from page one on every call, and you're billed per word. Perseus hands them just the page they need: it resolves live workspace state into verified facts before the context window opens, and pairs with Perseus Vault for durable, encrypted memory. The result: 73.8% on LongMemEval, a 67% smaller tool schema, and 611× warmer renders. Local-first, air-gap ready, MIT.

• 611×cold→warm cache speedup: 619.1s → 1.014s across 22,528 varied directives on the Perseus repo itself
• 171→57tool schema advertised once instead of tripled: a smaller payload every call
• 1.0semantic equivalence: 20/20 A/B pairs judged equivalent; same answers, fewer tokens

Persistent, encrypted memory for AI agents: one Rust binary, one file, no cloud.

Give your agents memory that survives the session, so they stop re-deriving what they already learned and stop repeating past mistakes. Hybrid recall (BM25 + dense + RRF), bi-temporal history, and AES-256-GCM at rest, exposed as 55+ MCP tools that work with any host. 73.8% on LongMemEval's official harness (vs Zep 63.8%, Mem0 49.0%). Local-first, air-gap ready, MIT.

• 73.8%LongMemEval QA, official harness (longmemeval_s 500), vs Zep 63.8%, Mem0 49.0%
• 3.7×faster hybrid recall: 295.5 → 80.9 ms p50 at 100K entities
• 98K/s100K entities inserted in 1.01s; ~85 MB RSS; recall unaffected under 100% GPU load