Tool Reference

Store decisions, facts, code patterns, and checkpoints.

add_memorysave_contextjournal_recordslicer_slice

Retrieve the right context without rereading the whole project.

query_memorysearch_memoryrecent_memorybrain_recall_episode

Turn repeated work into reusable procedures and corrected beliefs.

reflect_memoryjournal_distillbrain_consolidate_sleepbrain_reinforce

Use team and commons knowledge without mixing private namespaces.

commons_resonancebrain_discoverbrain_associatebrain_predict

Confirm auth, namespace, health, and usage before production work.

reunionnamespace_infomemory_statushealth_checkmnemosyne_context

Store a new memory entry with content and optional metadata.

When to use: When your AI learns something worth remembering — a pattern, decision, or fact.

Retrieve memories matching a natural-language query.

When to use: When your AI needs to recall specific knowledge from past sessions.

Full-text and semantic search across stored memories.

When to use: When you need broad search across everything stored — more exploratory than query.

Fetch the most recently stored memories in chronological order.

When to use: When you need context from the current or last session.

Trigger a reflection pass over stored memories to surface insights.

When to use: After a series of tasks — helps the AI distill patterns from accumulated experience.

Write a structured journal entry with mode, actor, outcome, and optional reward signal.

When to use: To log decisions, outcomes, and learning moments during a work session.

Search journal entries by content, mode, or time range.

When to use: To recall past decisions, search for how a problem was solved before.

Distill journal entries into consolidated learnings — observation-aware.

When to use: After a series of sessions — compresses journal history into reusable knowledge.

Discover related concepts and connections in the knowledge graph.

When to use: When exploring what the AI knows about a topic and its relationships.

Create an explicit association between two concepts in the graph.

When to use: When you want the AI to permanently link two ideas or patterns.

Recall a specific episode — a complete context window from a past session.

When to use: When you need the full context of a past interaction, not just a snippet.

Run a sleep-like consolidation pass — strengthens strong memories, prunes weak ones.

When to use: Periodically or at end-of-day — mimics how biological memory consolidates during sleep.

Apply reinforcement to specific memories based on outcome signals.

When to use: After completing a task — reward successful patterns, weaken unsuccessful ones.

Generate predictions based on learned patterns and associations.

When to use: When you want the AI to anticipate what comes next based on past experience.

Checkpoint the current working context for retrieval in future sessions.

When to use: Before switching tasks or ending a session — saves your place.

Pack maximum relevant context into the token budget using the Mnemosyne protocol.

When to use: At session start — hydrates the AI with the most relevant memories for the task.

Search the shared _commons knowledge pool for community knowledge.

When to use: When you want to tap into shared intelligence — threat intel, industry news, resolved issues.

Check the health of the MemQ service and backend connectivity.

When to use: To verify the service is running and all backends are reachable.

Get current memory usage statistics — count, storage, namespace info.

When to use: To check how much memory you're using and what namespaces are active.

Get details about the current namespace — tier, scope, permissions.

When to use: To verify which namespace you're connected to and what access you have.

Session handshake — greeting, namespace resolution, and context continuity.

When to use: At the start of every session. This is the first tool to call.

Slice content into optimally-sized chunks for memory storage.

When to use: When ingesting large documents or codebases into memory.