Structured memory for AI agents. Quick database for you.

Store customer feedback and waitlist emails or any data just like you would in a normal database, alongside agent memories, prompts, and project context. Everything stays organized, so nothing gets lost or mixed together. Your agents can search and update it automatically, while you can browse and manage the same data yourself.

Early access for indie hackers building with agents.

What is KeepDB?

Imagine customer feedback coming from your app, waitlist signups from your landing page, prompts used by your agents, project plans written by your team, and notes scattered across chats and documents. You want all of it searchable, organized, and easy to access later.

KeepDB gives humans and agents a shared place to store and retrieve that context. It is not a replacement for PostgreSQL or MongoDB. Instead, it is built for quick-access data like customer feedback, waitlist emails, prompts, documentation, notes, project plans, and agent memory without everything turning into one giant memory blob.

How does it work?

Tell your agent where something belongs.

Save this launch plan under the plans folder
Save these meeting notes under the product folder
Save this prompt under the prompts folder

Or skip the conversation entirely and send data directly from your app. Perfect for things like waitlist signups, customer feedback, support requests, feature requests, or any quick pipeline where setting up a full backend feels like overkill.

await fetch("https://api.keepdb.dev/memory", {
  method: "POST",
  headers: {
    Authorization: "Bearer keep_sk_..."
  },
  body: JSON.stringify({
    collection: "feedback",
    content: "Dark mode would be nice"
  })
})

Later, let your agent turn raw data into insights.

What features are users requesting the most?
Summarize feedback from the last 7 days
What was our migration plan?

One database. Multiple worlds.

Store customer feedback in one folder, waitlist emails in another, project plans somewhere else, and agent memory in its own dedicated space. Everything stays separated, searchable, and easy to reason about.

The same data can be updated by your applications, searched by your agents, and managed by you from a single place.

Need a normal database table for customers, feedback, or waitlist signups? Use it that way. Need structured memory for your agents? Use the same system.

One database. Multiple folders.

feedback
├─ User wants dark mode
├─ Search feels slow
└─ Love the AI summaries

waitlist
├─ raj@example.com
├─ sam@example.com
└─ jane@example.com

plans
├─ Launch strategy
├─ Pricing notes
└─ Marketing ideas

prompts
├─ Onboarding email
├─ Support assistant
└─ Landing page copy

agent-memory
├─ User preferences
├─ Project context
└─ Important decisions

Ready to stop losing context?

Keep customer data, agent memory, prompts, notes, and project context organized in one place that both humans and agents can use.

Common Questions