Persistent memory graph for AI coding agents — semantic search, knowledge graph, and time-based decay over MCP
Drift inferred · capture-to-capture
No drift recorded — single capability capture; advisories appear once its surface changes.
tools
-
content
Text; no Synapto length limit
-
depth_layer
Max 20 characters
-
find_contradictions
Find memory pairs that disagree
-
forget
Soft-delete a memory
-
get_memories
Fetch complete content for multiple recalled memories
-
get_memory
Fetch the complete content and metadata for one recalled memory
-
graph_query
Walk the knowledge graph (N-hop)
-
list_entities
Browse known entities
-
maintain
Run decay and cleanup
-
memory_stats
View counts and distribution
-
memory_type
Max 20 characters
-
recall
Search memories by meaning
-
relate
Link two entities ("Hermes" --[produces]--> "agent.messages")
-
remember
Store a memory (entities and search vectors are created automatically)
-
summary
Max 255 characters
-
tenant
Max 100 characters
-
trust_feedback
Mark a memory as helpful or unhelpful
-
update_memory
Replace, append, or patch fields on an existing memory
this server hasn't been statically analyzed yet (npm/PyPI artifacts only).