Core Concepts

Understanding REMBR's architecture and capabilities

What is REMBR?

REMBR is a memory-as-a-service platform that provides persistent, searchable context storage for AI applications. It's designed to be the context layer for AI assistants, agents, and recursive language models (RLMs).

MCP-Native
First-class Model Context Protocol support - not just an API wrapper
Multi-Tenant SaaS
Full tenant isolation with Row-Level Security for enterprise use
Hybrid Search
Combines semantic vector search and full-text search for best results

Memory Categories

Original Categories (8)

General-purpose memory types

facts
Concrete information and data points
preferences
User preferences and settings
conversations
Conversation history and context
projects
Project-specific information
learning
Knowledge and insights learned
goals
Objectives and targets
context
Situational context
reminders
Future actions and reminders

RLM-Optimized (4)

For coding and technical work

patterns
Code patterns, architectural patterns, best practices
decisions
Technical decisions, trade-offs, architectural choices
workflows
Process flows, deployment procedures
insights
Analytical findings, performance insights

Projects & Contexts

Projects

Organize memories by workspace

Projects are workspaces for organizing related memories. Each project can be:

  • Personal - Private to you (like "My Coding Notes")
  • Shared - Visible to your team (like "Company Wiki")

Create projects via the dashboard or your AI can create them automatically when needed.

Contexts

Focused memory spaces for specific tasks

Contexts are logical groupings within projects:

  • A context can contain memories from multiple categories
  • Memories can belong to multiple contexts
  • Perfect for focused search and analysis
Example Use Case
Context: "Authentication System" might contain memories of category patterns (JWT implementation), decisions (why we chose OAuth), and workflows (login flow).

Search & Retrieval

REMBR uses hybrid search combining multiple techniques:

Semantic Search
Finds conceptually similar memories using vector embeddings
Full-Text Search
Matches exact words and phrases using full-text search
Graph-Aware
Follows relationships between memories for richer context
💡 Pro Tip
You don't need to worry about search modes - just ask your AI naturally! REMBR automatically uses the best search strategy for your query.
Ready to dive deeper?
Explore the MCP Tools Reference to see all available capabilities