Memory Spaces Guide

Info

Last Updated: 2026-01-08

Multi-tenancy and isolation with Cortex Memory Spaces.

Overview

Memory Spaces provide isolated memory contexts for different users, teams, or projects. They are automatically created when you first use a memorySpaceId - no explicit registration is required for basic usage. Memory spaces provide complete data isolation, ensuring that all memories, conversations, facts, and contexts are scoped to a specific memory space.

Use Cases

1. Per-User Isolation

import { createCortexMemory } from "@cortexmemory/vercel-ai-provider";

const cortexMemory = createCortexMemory({
  convexUrl: process.env.CONVEX_URL!,
  memorySpaceId: `user-${currentUser.id}`,
  userId: currentUser.id,
  agentId: 'my-assistant',
});

2. Team Workspaces

import { createCortexMemory } from "@cortexmemory/vercel-ai-provider";

const cortexMemory = createCortexMemory({
  convexUrl: process.env.CONVEX_URL!,
  memorySpaceId: `team-${currentTeam.id}`,
  userId: currentUser.id,
  agentId: 'my-assistant',
});

3. Project-Based

import { createCortexMemory } from "@cortexmemory/vercel-ai-provider";

const cortexMemory = createCortexMemory({
  convexUrl: process.env.CONVEX_URL!,
  memorySpaceId: `project-${projectId}`,
  userId: currentUser.id,
  agentId: 'my-assistant',
});

Benefits

Data Isolation

Complete separation of data between memory spaces

Multi-Tenant SaaS

Built-in support for SaaS multi-tenancy

Team Collaboration

Share memory within teams or organizations

Per-Space Analytics

Track usage and performance per memory space

Managing Memory Spaces

For advanced usage, you can explicitly manage memory spaces using the Cortex SDK. Memory spaces can be registered, queried, updated, and deleted through the Memory Space Operations API. This includes capabilities like:

• Explicitly registering memory spaces with metadata
• Querying memory space details and statistics
• Updating memory space configurations
• Deleting memory spaces (with cascade options)

Next Steps