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

# What is Deepline?

> Deepline is the execution layer for repeatable GTM work done by coding agents.

Deepline is the execution layer for GTM work done by coding agents. You state
the outcome. Your agent selects or writes a Play. Deepline runs the Play and
keeps the inputs, outputs, and history for every run.

Use Deepline to enrich people or companies, research accounts, react to
provider events, and update connected business systems.

## Start with the interface you use

<CardGroup cols={2}>
  <Card title="Coding agent" icon="terminal" href="/docs/quickstart">
    Install Deepline in Claude Code, Codex, Cursor, or another agent that can
    use a shell. Then run one GTM request.
  </Card>

  <Card title="Claude plugin" icon="plug" href="/docs/claude-plugins">
    Add the Deepline plugin if Claude Code or Claude Cowork is where you work.
  </Card>

  <Card title="MCP client" icon="network" href="/docs/mcp/setup">
    Connect Deepline to ChatGPT Work, Claude Desktop, or another MCP client.
  </Card>

  <Card title="SDK" icon="code" href="/docs/sdk-v2/quickstart">
    Inspect, pilot, and scale a versioned workflow from your codebase.
  </Card>
</CardGroup>

## The three primitives

Skills give an agent the operating instructions to use these primitives.

### Plays

A **Play** is versioned code for repeatable GTM work. It runs inside Deepline.
Your agent writes or chooses the Play. Deepline records every run.

[Learn how Plays work](/docs/plays/plays-overview).

### Customer DB

**Customer DB** is your workspace's PostgreSQL database for customer records
and supported event tables. Its records persist across Play runs.

[Learn how to query Customer DB](/docs/database-access).

### Monitors

A **Monitor** pushes events from a supported external source into Deepline. It
writes accepted events to Customer DB. A connected Play can then enrich a
record, update a CRM, or alert a team.

[Learn how Monitors work](/docs/plays/monitors).

Plays execute code. Monitors ingest external events. Customer DB stores the
records they use. That is the whole system: receive a change, keep the data,
run code, and send the result where it belongs.

## How the system fits together

An agent can run a Play directly. A schedule, webhook, or Monitor event can
start a Play.

<Frame caption="How agents, data, events, and downstream systems connect.">
  <img src="https://mintcdn.com/deepline2/Sz2--KJIgGYAtP1_/images/deepline-system-architecture.svg?fit=max&auto=format&n=Sz2--KJIgGYAtP1_&q=85&s=9864fcdd7fc762125f6095552ba83304" alt="Deepline architecture: external sources push events through Monitors into Customer DB; Plays execute code inside Deepline; and Plays send actions to downstream systems." width="1504" height="560" data-path="images/deepline-system-architecture.svg" />
</Frame>

[Run your first Play](/docs/quickstart).