Skip to main content

Memory as a Service (MaaS)

LanOnasis Memory Suite provides a unified, enterprise-grade memory layer for applications and AI agents. It enables semantic search, persistent storage, topic organization, and full audit trails with compliance built-in (PCI-DSS, GDPR, SOX, HIPAA).

Platform Overview

Memory Suite acts as a central repository for structured information—project notes, user preferences, decision logs, AI context windows, session data, and compliance records. Unlike traditional databases, Memory Suite emphasizes:

  • Semantic Search: Find information by meaning, not just keywords
  • Multi-Interface Access: REST API, SDKs, CLI, MCP tools, or webhooks
  • Enterprise Compliance: Encryption, audit logging, retention policies, consent tracking
  • Real-Time Sync: Live updates across all connected clients and applications

Architecture

┌─────────────────────────────────────────────────────────────┐
│ LanOnasis Memory Suite │
├─────────────────────────────────────────────────────────────┤
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ REST API │ │ Memory CLI │ │ TypeScript │ │
│ │ (HTTP/REST) │ │ (Terminal) │ │ SDK │ │
│ └──────┬───────┘ └──────┬───────┘ └──────┬───────┘ │
│ │ │ │ │
│ └──────────────────┼──────────────────┘ │
│ │ │
│ ┌───────▼────────┐ │
│ │ MCP Tools │ │
│ │ (AI Agents) │ │
│ └───────┬────────┘ │
│ │ │
│ ┌──────────────────▼──────────────────┐ │
│ │ Memory Service (Core Engine) │ │
│ │ - Semantic Indexing (Embeddings) │ │
│ │ - Vector Search │ │
│ │ - Access Control (RBAC) │ │
│ │ - Audit Logging │ │
│ └──────────────────┬──────────────────┘ │
│ │ │
│ ┌───────────────────────┼───────────────────────┐ │
│ │ │ │ │
│ ┌──▼──────┐ ┌──────▼────────┐ ┌──────▼───┐ │
│ │ Storage │ │ Vector Index │ │ Audit │ │
│ │(Supabase) │ (pgvector) │ │ Log │ │
│ └─────────┘ └───────────────┘ └──────────┘ │
│ │
└─────────────────────────────────────────────────────────────┘

Core Capabilities

1. Structured Memory Storage

Store any structured data with automatic indexing:

{
"id": "mem_abc123",
"namespace": "projects/acme",
"text": "Q2 2026 product roadmap: AI features (priority), Performance optimization, Mobile app",
"tags": ["roadmap", "product-planning", "Q2-2026"],
"metadata": {
"owner": "alice@example.com",
"department": "Product",
"status": "approved"
},
"createdAt": "2026-01-15T10:30:00Z",
"updatedAt": "2026-01-15T10:30:00Z"
}

Search by meaning, not keywords. Find relevant memories even with different phrasing:

# Search for Q2 planning—will match all planning-related memories
lanonasis memory search --query "Q2 planning goals" --namespace projects/acme

Results include:

  • Exact keyword matches
  • Semantic matches (embeddings-based)
  • Tag-based filtering
  • Custom metadata filtering

3. Topic-Based Organization

Group related memories into topics for collaborative workflows:

# Create a topic for Q2 planning
lanonasis topic create \
--name "Q2 2026 Execution" \
--description "All planning, execution, and review for Q2"

Topics enable:

  • Collaborative organization (multiple team members)
  • Export as Markdown/PDF for reports
  • Comments and real-time updates
  • Version control of topic snapshots

4. Multi-Interface Access

Access memories through your preferred interface:

InterfaceUse CaseExample
REST APIBackend services, webhooks, integrationsPOST /api/v1/memories
CLITerminal workflows, CI/CD pipelines, scriptinglanonasis memory search --query "..."
TypeScript SDKNode.js/Bun applications, type safetymemory.create({...})
MCP ToolsAI agent context, automated reasoningmemory.search(query) via Claude/GPT
WebhooksReal-time event streamingReceive updates on create/update/delete

5. Enterprise Compliance

Built-in compliance features for regulated industries:

  • Encryption: AES-256-GCM for data at rest and in transit
  • Access Control: Role-based access control (RBAC) per namespace
  • Audit Logging: Immutable logs of all operations (PCI-DSS, SOX, HIPAA)
  • Data Retention: Automatic TTL and retention policy enforcement
  • GDPR: Right to deletion, consent tracking, data export
  • Consent Management: Mark sensitive memories as requiring explicit consent

Feature Comparison vs. Traditional Solutions

FeatureMemory SuiteTraditional DBVector DBNote-Taking App
Semantic Search✅ Native❌ No✅ Yes⚠️ Limited
Compliance (Audit, Encryption)✅ Built-in❌ Manual setup⚠️ Limited❌ None
Multi-Interface (REST, CLI, SDK, MCP)✅ All⚠️ REST only⚠️ Python mostly❌ Web UI only
Real-Time Sync✅ Yes⚠️ Polling⚠️ Limited❌ No
Namespace Isolation✅ Yes✅ Yes❌ No❌ No
TTL/Retention Policies✅ Yes⚠️ Manual❌ No❌ No
Golden Contract Compliance✅ Yes❌ No❌ No❌ No

Integration Patterns

Pattern 1: Application Context Storage

Store application state, user preferences, and session data:

// Node.js example
import { Memory } from "@lanonasis/memory-sdk";

const memory = new Memory({
apiKey: process.env.LANONASIS_API_KEY,
namespace: "app-context",
});

// Store user session
await memory.create({
text: "User session: alice@example.com, logged in from 10.0.0.5",
tags: ["session", "login"],
metadata: { userId: "u_123", timestamp: Date.now() },
});

Pattern 2: AI Agent Context Window

Provide persistent memory for AI agents to reason over:

// Claude or other AI model with Memory Suite integration
const relevantMemories = await memory.search({
query: "Q2 2026 goals and constraints",
namespace: "ai-context",
limit: 10,
});

// Feed to Claude
const response = await claude.messages.create({
model: "claude-3-5-sonnet",
system: `You are an AI assistant with access to organizational memory.\n\n${relevantMemories.map((m) => m.text).join("\n")}`,
messages: [{ role: "user", content: userQuery }],
});

Pattern 3: Compliance Audit Trail

Maintain immutable records of decisions and approvals:

lanonasis memory create \
--namespace compliance-audit \
--text "Security review: All PII encryption validated. Status: APPROVED" \
--metadata '{
"reviewer": "security@example.com",
"timestamp": "2026-01-15T14:30:00Z",
"reviewType": "security-audit",
"evidence": "security-report-2026-01-15.pdf"
}' \
--tags compliance,audited,pii-review

Pattern 4: Real-Time Team Collaboration

Team members synchronize context in real-time:

# Team lead creates a planning topic
lanonasis topic create --name "Sprint 42 Planning"

# Team members add memories and see updates in real-time
lanonasis topic add-memory \
--topic-id topic_xyz \
--memory-id mem_abc

# All team members subscribed to topic receive updates
# (via webhooks or real-time SDK)

Deployment Options

Pre-hosted at https://api.lanonasis.com

  • ✅ Fully managed infrastructure
  • ✅ Automatic scaling and high availability
  • ✅ Built-in compliance and security
  • ✅ No ops overhead

Option 2: Self-Hosted (Enterprise)

Deploy Memory Suite in your own infrastructure:

# Docker Compose setup
docker-compose up -d

# Configure with your database
export DATABASE_URL="postgresql://..."
export VECTOR_DB_URL="postgresql://..." # for pgvector

npm run migrate
npm run start

Requirements:

  • PostgreSQL 13+ with pgvector extension
  • Node.js 18+ runtime
  • Kubernetes recommended for production

Option 3: Hybrid (On-Prem with SaaS)

Critical data on-premise, non-sensitive data in SaaS:

// Route sensitive data to on-prem
if (isSensitive(memory)) {
await onPremMemory.create(memory);
} else {
await saasMemory.create(memory);
}

// Search across both
const results = await Promise.all([
onPremMemory.search(query),
saasMemory.search(query),
]);

Performance Characteristics

Latency

OperationLatency (p50)Latency (p99)Notes
Create50ms200msIncludes embedding generation
Search100ms500msSemantic search with 10 results
Get20ms100msBy ID (fastest)
Update75ms250msRe-indexes on metadata/tags
Delete30ms150msSoft delete by default

Throughput

  • Create: 1,000 memories/second per namespace
  • Search: 500 queries/second per namespace
  • Concurrent users: 10,000+ per deployment

Storage

  • Memory item: ~2-5 KB (raw) + 1 KB (embeddings)
  • Namespace quota: 10 GB default (configurable)
  • Retention: Unlimited (with optional TTL)

Security & Compliance

Encryption

  • In Transit: TLS 1.3
  • At Rest: AES-256-GCM with customer-managed keys (enterprise only)
  • Credential Storage: Bcrypt for passwords, AES-256-GCM for tokens

Access Control

  • Authentication: OAuth 2.0 (device flow, code flow, refresh tokens)
  • Authorization: RBAC per namespace; can delegate to external identity provider
  • Scope Management: Granular API key scopes (memory:read, memory:write, etc.)

Compliance Certifications

  • PCI-DSS 3.2.1: For payment card data
  • GDPR: With DPA and data processing agreements
  • SOX: Audit logging for financial controls
  • HIPAA: For health information (PHI)

Golden Contract v0.1: All operations follow Golden Contract v0.1 standards for credential rotation, token freshness, and audit compliance.


Getting Started

Quick Start (5 minutes)

# 1. Install CLI
npm install -g @lanonasis/cli

# 2. Authenticate
lanonasis auth device

# 3. Create a memory
lanonasis memory create \
--namespace default \
--text "My first memory!" \
--tags example

# 4. Search
lanonasis memory search --query "first" --namespace default

Next Steps

  1. CLI Documentation – Terminal commands and workflows
  2. REST API Reference – HTTP endpoints
  3. TypeScript SDK Guide – Programmatic access
  4. Authentication Guide – OAuth 2.0 flows
  5. Advanced Topics – Retention policies, webhooks, MCP integration

Use Cases

Project Management

Store project goals, decisions, constraints, and progress updates—searchable across your organization:

Project: Acme Q2 2026
├─ Goals: AI features, Performance
├─ Constraints: 3 engineers, $50K budget
├─ Decisions: Use Next.js 15, Prisma ORM
└─ Progress: Week 1–2 complete, Week 3 in progress

AI Agent Memory

Provide AI agents with persistent context to make better decisions:

Agent Context: Email Assistant
├─ User preferences: Concise responses, casual tone
├─ Company policies: Response SLA < 2 hours, escalate issues
├─ Recent interactions: Last 20 emails with summaries
└─ Decision log: Why certain emails were delegated

Compliance Auditing

Maintain immutable audit trails of decisions, reviews, and approvals:

Audit Trail: PII Processing
├─ 2026-01-10: Data inventory completed → APPROVED
├─ 2026-01-12: Encryption audit completed → APPROVED
├─ 2026-01-15: Access control review completed → APPROVED
└─ 2026-02-01: Quarterly compliance review scheduled

Team Collaboration

Team members share notes, decisions, and context in real-time:

Team: Product Design
├─ Sprint 42 Planning: Goals, design decisions
├─ Customer feedback: Latest requests and trends
├─ Competitive analysis: Market research notes
└─ Design system updates: Latest component changes


Support