Structured data,
built for AI agents.
Persistent databases, flexible schemas, relationships and files.
A data layer your agent can work with, without SQL setup.
How it works
Let your agent organize the data.
From a task to a useful structure.
Here’s a simple contacts example.
01 — DEFINE
Give the data a shape.
Start with a contacts schema. Define the fields your agent needs for the task.
Explore the example using the tabs.
| Field | Purpose | Example value |
|---|---|---|
name | Contact name | Alex Chen |
email | Email address | alex@example.com |
company | Related organization | Northwind |
notes | Useful context | Interested in the alpha |
| Name | Company | |
|---|---|---|
| Alex Chen | alex@example.com | Northwind |
| Sam Rivera | sam@example.com | Fieldwork |
| Robin Park | robin@example.com | Northwind |
A small set of records, organized around the same schema.
Connect contacts to companies rather than repeating information.
Capabilities
The pieces your agent needs.
Structured storage and retrieval,
in one place.
- Persistent data
- A database that gives your agent a place to keep structured information.
- Storage
- Flexible schemas
- Define fields around the task and give records a consistent structure.
- Structure
- Relationships
- Connect related records so the information stays useful in context.
- Connections
- Files
- Work with files alongside the structured information your agent manages.
- Content
- Querying
- Retrieve the information needed for the next step in your application.
- Retrieval
- MCP · API · CLI
- Interfaces in the alpha direction. Connection details are shared with participants.
- Integration
Getting started
Start with the private alpha.
AgentDB has a working v0.1 alpha release candidate. We’re inviting developers to explore it while the product takes shape.
Getting started guide- 1
Tell us your use case
Describe the agent or application you want to build.
- 2
Receive an invitation
Access is by invitation during the private alpha.
- 3
Follow the alpha setup guide
Use the connection details provided with your invitation.
Early access
What are you building?
Tell us how you’d use AgentDB.
A few sentences about your workflow are enough.
Private alpha · v0.1
Access by invitation