Skip to main content

Nest/Branch Organization and Agent Isolation

Nest/Branch Hierarchy

Organize memories into a two-level taxonomy. Nests are top-level domains (projects, people, topics). Branches are subtopics within nests (auth, billing, deployment).

ToolWhat it does
localnest_nest_listList all nests with memory counts
localnest_nest_branchesList branches within a specific nest
localnest_nest_treeFull taxonomy tree: nests, branches, counts

Pass nest and branch parameters when storing memories for organized retrieval. Use them as filters on localnest_memory_recall for focused results.

Agent Isolation

Multiple AI agents can share the same LocalNest instance without seeing each other's private data.

  • agent_id on memory entries enables per-agent scoping
  • Recall shows own memories + global memories, never another agent's private entries
  • Diary is a separate private scratchpad per agent
ToolWhat it does
localnest_diary_writeWrite a private diary entry (agent-isolated)
localnest_diary_readRead your own recent diary entries

Conversation Ingestion

Import past conversations into structured memory with automatic entity extraction.

ToolWhat it does
localnest_ingest_markdownImport Markdown conversation exports
localnest_ingest_jsonImport JSON conversation exports
localnest_memory_check_duplicateCheck for semantic duplicates before filing

Ingestion splits conversations into per-turn memory entries, extracts entities via rule-based heuristics, creates KG triples, and runs dedup. Re-ingestion of the same file is skipped by content hash.

CLI

localnest ingest ./chat-export.md --nest conversations --branch project-x
localnest ingest ./export.json --format json