> ## Documentation Index
> Fetch the complete documentation index at: https://hyperfx.ai/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Hyper Database

> Give agents structured, queryable storage. Use managed tables Hyper auto-creates from agent work, or define your own custom tables.

**Hyper Database** is the structured-data layer for your agents. Where [Knowledge Bases](/data/knowledge-bases) cover unstructured docs, Hyper Database handles tables — rows of records that agents can read, write, query, and update.

Two flavors:

* **Managed tables** — Hyper auto-creates these as agents do their work (e.g. a "leads" table that fills up when your prospecting agent runs, a "campaigns" table that builds up over time)
* **User-defined tables** — you create the schema yourself; agents read and write against it

Both are queryable from any agent in your workspace.

## When to use this

Use Hyper Database when:

* The data is **structured** (rows and columns, not docs)
* Multiple agents and tasks need to **read and write to the same source of truth**
* You want **historical state** an agent can update over time, not just one-shot context
* You're building a **custom workflow** where the agent's output feeds the next run's input

Examples:

* A leads table populated by a prospecting agent and dequeued by an outreach agent
* A campaigns table tracking spend, performance, and status across ad platforms
* A research-tracker table storing every competitor scan run and its findings
* A content calendar where a writing agent picks up ideas, drafts, and marks them shipped

## Managed tables

Some agent templates and skills automatically create and maintain managed tables for you. For example, the Apollo prospect skill writes lead records to a managed `leads` table; the platform-usage-analysis skill writes findings to its own table.

You don't need to set these up — they appear in **Hyper Database** as agents use them. You can read, query, and even export them like any other table.

## Custom tables

To define your own:

<Steps>
  <Step title="Open Hyper Database">
    From the sidebar, click **More** → **Hyper Database** → **Create table**.
  </Step>

  <Step title="Define the schema">
    Add columns and pick a type for each (text, number, date, boolean, JSON). Hyper handles the storage layer.
  </Step>

  <Step title="Seed it (optional)">
    Import a CSV or paste rows to start with existing data, or leave it empty for agents to populate.
  </Step>

  <Step title="Attach to an agent">
    Open an agent's **Resources** tab and add the table. The agent can now query it, append rows, update existing rows, and run aggregations.
  </Step>
</Steps>

## Queries from chat

Agents handle SQL-like queries themselves. You ask in plain language:

```text theme={null}
From the leads table, find all contacts in the SaaS vertical that we
haven't emailed in the last 30 days and rank them by company headcount.
Save the result as a new view.
```

The agent figures out the query, runs it, and returns results — and can chain it into the next action (draft outreach, push to a sequence, etc.).

## Powering tasks

Database tables are most powerful inside [tasks](/agents#tasks). For example:

* A daily prospecting task pulls 50 new leads into the `leads` table
* An hourly outreach task picks the top 10 unworked leads, drafts personalized emails, and updates the row status
* A weekly reporting task aggregates the table and posts a summary to Slack

The table is the shared state across all of those runs.

## Going further

* Build a [task](/agents#tasks) that reads or writes to your database on a schedule
* Visualize tables with [Interfaces](/data/interfaces) — turn them into shared dashboards
* For unstructured docs, use [Knowledge Bases](/data/knowledge-bases) instead