Swapnanil Saha

An AST-aware codebase indexer with hybrid semantic + BM25 search, a symbol graph, and a persistent working-memory tier delivered through hooks — findings survive context compaction and session restarts instead of dying with the conversation. Any MCP-compatible AI code editor connects automatically. v1.1 added team mode — many agents, one shared memory; v1.3 added per-memory triggers, so a note resurfaces exactly when its path, symbol, or lifecycle condition matches; v1.5 added automatic episode capture, where the harness records what was tried and what failed, and distils the recurring arcs into notes; v1.8 added a machine-checked user-stated trust class, where a note quoting the user's own words verbatim proves it; v1.10 made expiry a state transition instead of a delete, so an aged-out note is marked expired and stays auditable rather than vanishing, added pinned notes that recall returns whatever the query says, and rendered working memory to a read-only MEMORY.md a person can actually open and read; v1.11 let a note be retired after the fact without the "proven wrong" framing revocation carries, recorded whether the agent actually read the file a note is anchored to rather than assuming it, and opened working memory in your own editor while still writing every change back through the append-only lifecycle; v1.12 gives memory an inbound path, so an existing MEMORY.md, CLAUDE.md or .cursorrules can be imported into the store instead of retyped, with imported notes marked as imported rather than passed off as things the agent worked out itself.

v1.12 gives memory a way in. vectr could always write memory out and never read it back, so adopting it began by retyping whatever you had already written down. vectr memory import reads an existing MEMORY.md, CLAUDE.md, AGENTS.md or .cursorrules into the store, previews the whole plan before writing anything, and marks what it imported as imported, because a rule read off a file is a guess about what you meant and should not arrive wearing the authority of something the agent verified. The release also found a similarity metric that had been reading almost the inverse of the truth on repetitive code, which had been quietly holding a dedup key together. v1.11 separated being out of date from being wrong: a note can now be retired after the fact in favour of a newer one, without the "proven wrong" framing revocation carries, and reinstating one actually reverses it instead of leaving it half-retired. An anchor now records whether the agent ever actually read the file it points at, so a note anchored to a file nobody opened no longer looks identical to one that was genuinely bound. And working memory opens in your own editor, with every change written back through the append-only lifecycle rather than around it. v1.10 made memory expiry a state transition rather than a delete: an aged-out note is marked expired and keeps its place in the audit trail, directives never age out at any age, and a pinned note comes back from recall whatever the query says. It also rendered a workspace's notes to a read-only MEMORY.md, so what the agent remembers is finally something a person can open and read. v1.8 shipped a machine-checked user-stated provenance class (a note binding a verbatim excerpt of the user's words earns a stronger recall frame; the class is earned by a deterministic check, never claimed), after 1.6 and 1.7 made proactive delivery the default and made it worth reading. v1.5 shipped automatic episode capture with arc distillation, and a note lifecycle where a wrong note is revoked rather than deleted, so the mistake is not silently re-derived, on top of 1.3's per-memory trigger engine and 1.1's team mode. The research underneath it is now published: Delivery, Not Storage (arXiv:2607.20972), a controlled study of agent memory under repeated context compaction across 108 and 138 forced boundaries, graded token by token. A second report from the same programme is published beside it: Deterministic, Reproducible, and Wrong, on how our own deterministic scorer manufactured an agent-memory headline that the transcripts refuted, and the contradiction guard that would have caught it. Running alongside it: an upstream feature request to Claude Code (#78795) proposing triggered injection for auto-memory files, so the findings become platform behavior rather than a workaround.