> ## Documentation Index
> Fetch the complete documentation index at: https://docs.e6data.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Before you begin

> What must be in place before anyone can use AgentFlow: a running deployment, authentication, a model, an e6data cluster and storage.

export const DocMeta = ({type, applies, owner}) => {
  const LABELS = {
    tutorial: "Tutorial",
    "how-to": "How-to guide",
    explanation: "Explanation",
    reference: "Reference"
  };
  return <div className="e6-meta">
      <span className={type === "how-to" || type === "tutorial" ? "e6-eyebrow" : "e6-eyebrow e6-eyebrow--quiet"}>{LABELS[type] || type}</span>
      {applies ? <span>Applies to: {applies}</span> : null}
      {owner ? <span>Owner: {owner}</span> : null}
    </div>;
};

<DocMeta type="how-to" applies="All deployments · for administrators" />

Complete these once per deployment. Your users can't create a project or start a chat until they're done.

<Steps>
  <Step title="Run an AgentFlow deployment">
    Choose a deployment model: SaaS, BYOC, in-VPC, or on-premises and private cloud. See [Deployment models](/agentflow/operate/deployment-models).
  </Step>

  <Step title="Configure authentication">
    Use the e6data session and SSO path, or the standalone JWT path. You can also enable personal access tokens for machine access and session exchange.
  </Step>

  <Step title="Enable at least one model">
    Turn on a model in the Model Registry. Supported providers are Anthropic, OpenAI, Amazon Bedrock, OpenRouter, and OpenAI-compatible proxy endpoints. See [Models](/agentflow/operate/models).
  </Step>

  <Step title="Make an e6data cluster available">
    AgentFlow needs an active e6data cluster to build semantic layers and run SQL. See [Query Engine clusters](/query/clusters/create-and-manage).
  </Step>

  <Step title="Configure durable storage and Redis">
    Durable storage holds workflows, skills, history and audit records. Redis coordinates live work. See [Architecture](/agentflow/concepts/architecture).
  </Step>

  <Step title="Turn on optional features you need">
    Uploads, external Cube, Slack, GitHub, GreptimeDB, MCP, enrichers, actions, triggers and response contracts are each enabled separately.
  </Step>
</Steps>

## Next

<Columns cols={2}>
  <Card title="Ask your first question" icon="message-square" href="/agentflow/get-started/first-answer">Create a project and a semantic layer, then chat with your data.</Card>
  <Card title="Deployment models" icon="server" href="/agentflow/operate/deployment-models">Pick where AgentFlow runs.</Card>
</Columns>


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.