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

# Deep Agents Code

> Terminal coding agent built on the Deep Agents SDK

Deep Agents Code (`dcode`) is an open source coding agent built on the [Deep Agents SDK](/oss/javascript/deepagents/quickstart).
It works with any large language model and supports switching providers or models.
Persistent memory carries context across conversations, customizable skills shape behavior, and approval controls gate code execution.

## Get started

Run the following command to install Deep Agents Code and launch an interactive session:

```bash theme={"theme":{"light":"catppuccin-latte","dark":"catppuccin-mocha"}}
curl -LsSf https://langch.in/dcode | bash
dcode
```

See the [Quickstart](/oss/javascript/deepagents/code/quickstart) to add provider credentials, run your first task, and learn interactive mode.

<Frame>
  <video autoPlay muted loop playsInline className="w-full aspect-video rounded-xl" src="https://mintcdn.com/langchain-5e9cc07a/RVTbVyxmLiI04cgS/oss/images/deepagents/dcode-small.mp4?fit=max&auto=format&n=RVTbVyxmLiI04cgS&q=85&s=0d35e29a34f349183e83bd3d1eceb68b" aria-label="Deep Agents Code terminal demo" data-path="oss/images/deepagents/dcode-small.mp4">
    Your browser does not support the video tag.
  </video>
</Frame>

## Capabilities

<CardGroup cols={3}>
  <Card title="Remote sandboxes" icon="cloud" href="/oss/javascript/deepagents/code/remote-sandboxes">
    Run agent tools remotely instead of on your local machine.
  </Card>

  <Card title="Goals and rubrics" icon="target-arrow" href="/oss/javascript/deepagents/code/goals-and-rubrics">
    Define measurable objectives or grading criteria so the agent can check whether work is done.
  </Card>

  <Card title="Subagents" icon="users" href="/oss/javascript/deepagents/code/subagents">
    Delegate work to task-specific subagents for parallel execution.
  </Card>

  <Card title="Memory" icon="brain" href="/oss/javascript/deepagents/code/memory-and-skills#memory">
    Store and retrieve information across sessions, including project conventions and learned patterns.
  </Card>

  <Card title="Context compaction" icon="arrows-minimize" href="/oss/javascript/deepagents/code/quickstart#interactive-mode">
    Summarize older messages and offload originals to storage.
  </Card>

  <Card title="Human-in-the-loop" icon="user" href="/oss/javascript/deepagents/code/quickstart#interactive-mode">
    Require human approval for sensitive tool operations.
  </Card>

  <Card title="Skills" icon="puzzle" href="/oss/javascript/deepagents/code/memory-and-skills#skills">
    Extend agent capabilities with custom expertise and instructions.
  </Card>

  <Card title="MCP tools" icon="plug" href="/oss/javascript/deepagents/code/mcp-tools">
    Load external tools from Model Context Protocol servers.
  </Card>

  <Card title="Tracing" icon="chart-dots" href="/oss/javascript/deepagents/code/quickstart#trace-with-langsmith">
    Trace agent operations in LangSmith for observability and debugging.
  </Card>
</CardGroup>

## Next steps

<CardGroup cols={2}>
  <Card title="Quickstart" icon="player-play" href="/oss/javascript/deepagents/code/quickstart">
    Install Deep Agents Code, run your first task, and use interactive or non-interactive modes.
  </Card>

  <Card title="Configuration" icon="settings" href="/oss/javascript/deepagents/code/configuration">
    Set up credentials, `config.toml`, environment variables, hooks, and CLI flags.
  </Card>
</CardGroup>

***

<div className="source-links">
  <Callout icon="terminal-2">
    [Connect these docs](/use-these-docs) to Claude, VSCode, and more via MCP for real-time answers.
  </Callout>

  <Callout icon="edit">
    [Edit this page on GitHub](https://github.com/langchain-ai/docs/edit/main/src/oss/deepagents/code/overview.mdx) or [file an issue](https://github.com/langchain-ai/docs/issues/new/choose).
  </Callout>
</div>