Jan 12 2026 3:15 PM
Prefers TypeScript with strict mode enabled
Saved from Claude
One Memory for All Your AI Tools
Tell one AI, and the rest will know
One shared memory layer
See it in Action:
Built for people who use multiple AI tools
Reflect Memory gives your AI tools shared context so they work better together.
Integrations
7+
ChatGPT, Claude, Cursor, Gemini, Grok, Lovable, n8n
Recall Speed
< 50ms
Real-time recall across all connected tools
Memory Types
3
Semantic, episodic, and procedural
Privacy
100%
You own your data. We never train on it.
Works with the AI tools you already use
How it works
Save something in one AI tool, and every other connected tool can access it. No lock-in, no complexity. Your data stays yours.
Feb 3 2026 10:42 AM
Launched v2 API with OAuth 2.1 support
Saved from ChatGPT
Mar 8 2026 5:20 PM
Deploy to Railway, use Vercel for frontend
Saved from Cursor
Accurate by design
No AI guessing what to remember. What you store is exactly what gets recalled. Every update is versioned so you can always see the full history.
Connects in minutes
Add Reflect Memory to your AI tools and they automatically discover how to read, write, and search your memories. No complicated setup required.
Works with ChatGPT, Claude, Cursor, and more out of the box.
1
ConnectAdd your AI tool
2
SaveMemories sync instantly
3
RecallAny tool, any time
What changes when your AIs share memory
Second opinions with full context
Run a strategy through Claude. Hand it to ChatGPT. It already knows everything and pushes back on the weak spots.
Decisions that stick across tools
Make a call in Cursor today. Open Claude tomorrow. It remembers what you decided, why, and what you ruled out.
AIs that build on each other
Use different tools for different strengths. They pick up where the last one left off instead of starting over.
Security and privacy, by default
Zero data training
We never train AI models on your data. Your memories are yours. Period.
Encrypted in transit
All API communication is encrypted via TLS. Memories are scoped to your account with timing-safe authentication.
Per-user memory isolation
Each user's memory store is private to them. No cross-account data leakage by design.
Full audit trail
Every read, write, and auth event is logged. Queryable via API and exportable for compliance reviews.
SSO and enterprise auth
Integrate with Okta, Azure AD, Google Workspace, Auth0, or Keycloak via OIDC. PKCE-based OAuth for AI tool connectors.
No telemetry, no phone-home
Private deployments run fully air-gapped. No usage data, analytics, or telemetry leaves your network.
Cloud or private deploy. Your choice.
Use our hosted cloud for instant setup, or deploy privately in your own infrastructure for full data control.
Need data to stay in your network?
Private deploy is a single Docker container with env-based configuration. We'll have you running in 30 minutes.
Simple, transparent pricing
Start free. Upgrade when your AI workflow demands it.
Free
For individuals exploring AI memory
$0
Free forever- 500 memories
- 2 API keys
- All integrations included
- 1 user
- Memory versioning
- Community support
Pro
For power users who need more capacity
$20
per month- 5,000 memories
- 5 API keys
- All integrations included
- 1 user
- Priority support
- Everything in Free
Most Popular
Team
Shared memory for your whole team
$50
per month- 10,000 shared memories
- 10 API keys (team-wide)
- Up to 10 seats
- Shared namespace
- Memory attribution
- “Share to Team” controls
- Everything in Pro
Need more? Talk to us
Enterprise
Private deploy with full control
Custom
Tailored to your organization- Unlimited memories
- Private deploy (Docker / Helm)
- SSO / OIDC integration
- Full audit trail + export
- Model egress control
- Dedicated support + SLA
- Everything in Team
Frequently asked questions
Built for builders using AI every day
Start free. No credit card required. Connect your first AI tool in under two minutes.