About Memstrata
Memstrata is an MCP server published by ConsultRuss in the RAG category: semantic memory MCP server (FastMCP + Postgres/pgvector) with tiered recall depth for token economy. Ships a retrieval-quality eval suite — precision@5 0.69, recall@10 0.82.
Install
git clone https://github.com/ConsultRuss/memstrataThis server has no ConduID identity, so agent calls to it are not receipted. Pin the version you install and review the source before granting it credentials.
Tools (8)
delete_memoryIrreversible removal
Parameters not published.
get_briefingsActive briefings for session start
Parameters not published.
get_memoryFull record by ID
Parameters not published.
pin_briefingPin a memory as an active briefing with TTL
Parameters not published.
save_memoryStore content + optional summary, category, metadata; embeds on write
Parameters not published.
search_memoriesSemantic (pgvector cosine) or keyword (`websearch_to_tsquery`) via `mode`; depth via `depth`
Parameters not published.
statsCounts by category, embedding coverage, briefing slots
Parameters not published.
update_memoryMutate content/metadata — **re-embeds when content changes**
Parameters not published.
Ask AI
Ask AI about Memstrata
Powered by Claude · Grounded in docs
Security checks
6 checks have not been run yet.