RembrREMBR

MCP Tools Reference

All MCP Tools available to your AI

Core Memory (6 tools)

store_memory

Store a new memory with metadata and category

Parameters:
  • content (required) - The content to remember
  • category (required) - One of 12 categories
  • metadata (optional) - Additional key-value pairs
  • relevance_score (optional) - 0.0-1.0 relevance score

search_memory

Search memories using hybrid text and semantic search

Parameters:
  • query (required) - Search query
  • category (optional) - Category filter
  • limit (optional) - Max results (1-50, default 10)
  • min_similarity (optional) - Minimum score (default 0.7)
  • search_mode (optional) - "hybrid", "semantic", "text", "phrase"
  • metadata_filter (optional) - Filter by metadata fields (e.g., {"topic": "auth"})

get_memory, update_memory, delete_memory, list_memories

CRUD operations for managing memories

Standard operations for retrieving, updating, and deleting individual memories or listing recent ones.

Enhanced Memory & Search (5 tools)

find_similar_memories

Find memories similar to a specific memory using semantic similarity

Great for discovering related context you may have forgotten about.

filter_memories, batch_memories

Filter memories by category, date range, and metadata; batch store or update multiple memories at once

saved_searches

Save, list, and execute reusable search queries

get_stats, get_embedding_stats

Get memory statistics and usage information

View plan limits, usage percentages, and semantic search status.

Context Management (4 tools)

create_context, list_contexts

Create and manage context workspaces

Organize memories into focused workspaces for specific tasks or topics.

search_context

Search within a specific context using scoped hybrid search

More accurate results when you know the context you're working in.

add_memory_to_context

Link existing memories to contexts (many-to-many)

Memories can be in multiple contexts for flexible organization.

Snapshots & Temporal (10 tools)

create_snapshot, get_snapshot, list_snapshots

Create immutable snapshots for time-travel debugging

Perfect for understanding what an agent knew at a specific point in time.

create_temporal_snapshot, list_temporal_snapshots

Named snapshots for fast temporal queries

Create checkpoints at important moments for later analysis.

search_at_time, get_memory_history, compare_snapshots

Query memories as they existed at specific times

Debug decisions: "What did the agent know when it chose X?"

snapshot_timeline, snapshot_diff, snapshot_nearest, snapshot_category_evolution

Advanced snapshot analysis: timelines, diffs between snapshots, nearest-in-time lookup, and category evolution tracking

Graph Analytics (4 tools)

get_memory_graph

Get relationship graph for a context

Visualize how memories connect to each other.

detect_contradictions

Find contradicting memory pairs within a context

Identify inconsistencies in stored information.

get_context_insights, classify_query_intent

Get insights and classify query intent

Pre-compiled analytics and AI-powered query classification.

Context Intelligence (3 tools)

infer_memory_relationships

Infer semantic relationships for a specific memory

Automatically discover how memories relate to each other.

enhanced_search

Smart search with intent classification and relationship expansion

Get better results by understanding query intent and following relationships.

get_memory_insights

Get insights about memory patterns and relationships

Understand how a specific memory fits into your knowledge graph.

Analytics & Reporting (10 tools)

detect_memory_contradictions

Detect contradictions with confidence scoring

generate_context_graph

Generate interactive context graph with clusters

generate_memory_insights

Generate insights about memory usage patterns

get_predictive_analytics

Forecast memory growth and quality trends

get_usage_analytics, get_performance_metrics

Usage analytics and performance metrics across your tenant

get_memory_growth, get_category_breakdown

Track memory growth over time and breakdown by category

get_pii_analytics, get_context_insights

PII scan analytics and pre-compiled context insights

build_report, export_memories

Build custom reports and export memories in bulk

Personal Memory (2 tools)

set_memory_visibility

Change visibility scope (personal, shared, project)

list_personal_memories

List personal memories for current user only

Causal Reasoning (4 tools)

trace_causality

Trace causal reasoning chain from a memory

Critical for RLM debugging: understand why an agent made a decision.

infer_causality

Infer causal relationship between two memories

get_causal_links, validate_causal_link

Get and validate causal relationships

Audit, Compliance & GDPR (10 tools)

query_audit_log

Query audit logs with filters (action type, date range, actor)

generate_compliance_report

Generate SOC2/compliance audit report

get_audit_stats

Audit logging statistics and health metrics

audit_health, audit_metrics

Real-time audit system health and detailed metrics

audit_evaluate_thresholds, audit_alerts

Threshold evaluation and alert management for compliance

audit_anomaly_detect

Detect anomalous patterns in audit data

audit_metrics_prometheus

Export audit metrics in Prometheus format

gdpr

GDPR operations: forget-me, data export, consent management

PII Detection (4 tools)

Detect, redact, and score personally identifiable information across 21 entity types using the built-in NLP engine.

pii

Scan text for PII and return detected entities with positions

pii_nlp_detect

Advanced NLP-based PII detection with confidence scores

pii_nlp_redact

Redact PII from text, replacing with category placeholders

pii_nlp_score

Score text for PII risk level (low/medium/high/critical)

File Storage (5 tools)

Attach files to memories using MinIO-backed object storage with automatic content-type detection.

upload_attachment

Upload a file attachment linked to a memory

list_attachments

List all attachments for a memory

get_attachment_url

Get a pre-signed download URL for an attachment

delete_attachment

Delete an attachment from a memory

get_storage_usage

Get storage usage statistics for your tenant

⚙Task Management & RLM (10 tools)

Core infrastructure for Recursive Language Model workflows: task lifecycle, iteration tracking, stuck detection, inter-agent handoffs, and work queues.

manage_task

Task CRUD + assignment. Operations: create, get, update, delete, list, assign

task_state

State machine transitions (pending→in_progress→blocked/completed/failed) with history

task_dependencies

DAG dependency management with cycle detection and critical path analysis

task_iterations

Record execution attempts, calculate stuck score (0-100), detect stuck tasks across tenant

Stuck detection: plateau (+35pts), repeating errors (+25pts), high iteration count (+20pts), idle time (+20pts). Score ≥70 = stuck.

task_handoff

Inter-agent task transfer. Operations: create, accept, reject, list_pending, get, history

manage_acceptance_criteria

Per-task acceptance criteria with evidence linking. Operations: add, validate, status, link_evidence, list, delete

work_queue

Priority queue with Redis-style semantics. Supports bulk_enqueue, claim (race-safe), bulk_complete, stats

task_export

Export/import tasks with dependencies and acceptance criteria for RLM portability

task_analytics

Velocity (tasks/period + trend), burndown (remaining work + projection), bottlenecks (blocked/overloaded agents)

rlm_session + rlm_iteration + rlm_regenerate

Full RLM session lifecycle: create sessions, track iterations, evaluate acceptance criteria, regenerate stuck plans