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

# What's a project

A project groups the resources for one voice use case. Depending on your path,
it contains:

* An agent with its Listen, Think, and Speak configuration.
* Models, API keys, and secrets for the project.
* Tools and telephony for inbound and outbound calls.
* Call history and usage.

You can also use a project with the API and no agent. Resources you create
outside a project, such as BYOK keys, stay global and are available across
projects.

## Pick your path

When you create a project, you pick how SLNG fits into your stack. Each path
sets the project up differently, so choose the one that matches how you work.

<CardGroup cols={1}>
  <Card title="Deploy on SLNG" icon="rocket" href="/guides/get-started/create-a-project/create-agent">
    Create an agent that runs on SLNG. Configure it in the dashboard or through
    the API.
  </Card>

  <Card title="Improve an existing agent, in your own code (LiveKit / Pipecat / custom)" icon="code" href="/guides/get-started/create-a-project/improve-agent#in-your-own-code">
    Keep your agent in your codebase. Choose the integration for your framework
    or custom client.
  </Card>

  <Card title="Improve an existing agent, on a voice platform (ElevenLabs / Vapi / Deepgram)" icon="phone" href="/guides/get-started/create-a-project/improve-agent#on-a-voice-platform">
    Keep your agent on its current platform. Choose the platform to see its
    integration steps.
  </Card>

  <Card title="Use the API directly" icon="terminal" href="/guides/get-started/create-a-project/api-only">
    Call SLNG models straight from the API, without building an agent.
  </Card>
</CardGroup>