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

# Installation Options for Tyk AI Studio

> Explore the different installation options for Tyk AI Studio, including Docker, Kubernetes, and Linux.

export const ResponsiveGrid = ({children}) => {
  return <div className="responsive-grid">
      {children}
    </div>;
};

## Availability

| Edition                                                                                                                                         | Deployment Type      |
| :---------------------------------------------------------------------------------------------------------------------------------------------- | :------------------- |
| [Community](/5.12/ai-management/ai-studio/overview#community-edition) & [Enterprise](/5.12/ai-management/ai-studio/overview#enterprise-edition) | Self-Managed, Hybrid |

## Deployment Models

Tyk AI Studio supports two deployment modes:

### Standalone

Standalone mode operates as a single AI Studio instance with embedded gateway functionality. It is suitable for development, testing, and small teams.

```mermaid theme={null}
graph LR
    Users[Users] --> Studio[AI Studio<br/>Standalone]
    Studio --> DB[(SQLite)]
    Studio --> Providers[AI Providers]
    
    style Studio fill:#4A90E2
```

**Characteristics**:

* Single instance deployment
* Built-in gateway functionality
* SQLite database (or PostgreSQL)
* No external dependencies

### Hub and Spoke (Control + Edge)

Control Plane with Edge Gateways uses AI Studio as the central control plane managing edge gateways for distributed request processing. This approach uses lightweight Edge Gateways instances that connect to the control plane.

It is suitable for production, enterprise, and multi-region deployments.

```mermaid theme={null}
graph TB
    subgraph "Control Plane"
        Control[AI Studio<br/>Control Mode]
        ControlDB[(PostgreSQL)]
        Control --> ControlDB
    end
    
    subgraph "Edge Gateways"
        Edge1[Edge Gateway 1]
        Edge2[Edge Gateway 2]
        EdgeN[Edge Gateway N]
    end
    
    Control -.->|Config Sync| Edge1
    Control -.->|Config Sync| Edge2
    Control -.->|Config Sync| EdgeN
    
    Users1[Users Region 1] --> Edge1
    Users2[Users Region 2] --> Edge2
    UsersN[Users Region N] --> EdgeN
    
    Edge1 --> Providers[AI Providers]
    Edge2 --> Providers
    EdgeN --> Providers
    
    style Control fill:#4A90E2
    style Edge1 fill:#7ED321
    style Edge2 fill:#7ED321
    style EdgeN fill:#7ED321
```

**Characteristics**:

* Centralized configuration management
* Distributed request processing
* Regional edge deployments
* High availability and fault tolerance
* Namespace-based multi-tenancy

## Choosing Your Deployment

| Scenario                   | Recommended Mode | Why                      |
| -------------------------- | ---------------- | ------------------------ |
| Local development          | Standalone       | Simple, fast setup       |
| Single office/location     | Standalone       | No distribution needed   |
| Multiple regions           | Hub-and-Spoke    | Low latency for users    |
| High availability          | Hub-and-Spoke    | Fault tolerance          |
| Multi-tenant SaaS          | Hub-and-Spoke    | Namespace isolation      |
| Compliance (data locality) | Hub-and-Spoke    | Regional data processing |

## Requirements

**Tyk AI Studio** and **Edge Gateway** requires a persistent datastore for its operations. By default, SQLite is used, while PostgreSQL is recommended for production deployments.

### Required Components

| Component      | AI Studio (Hub)           | Edge Gateway (Edge)       |
| -------------- | ------------------------- | ------------------------- |
| **SQLite**     | ✅ **Supported** (default) | ✅ **Supported** (default) |
| **PostgreSQL** | ✅ **Supported**           | ✅ **Supported**           |

### Optional Components

**Tyk AI Studio** uses a message queue for chat interface. By default, an in-memory queue is used, which is suitable for development and single-instance deployments.

For production and multi-instance deployments, you can configure AI Studio to use [NATS JetStream](/5.12/ai-management/ai-studio/installation/nats) as the message queue backend.

## Recommended Installation: Docker

For development, testing, and proof of concept purposes, we recommend using our Docker installation, which allows you to quickly spin up AI Studio on your local machine.

<ResponsiveGrid>
  <Card href="/5.12/ai-management/ai-studio/quickstart" img="https://mintcdn.com/tyk/zB4143fn76CY8N8G/img/docker.png?fit=max&auto=format&n=zB4143fn76CY8N8G&q=85&s=543bd6db2669aa77580d09d33ada2bfb" width="71" height="61" data-path="img/docker.png">
    Install with Docker
  </Card>
</ResponsiveGrid>

## Alternative Installation Methods

<ResponsiveGrid>
  <Card href="/5.12/ai-management/ai-studio/deployment-k8s" img="https://mintcdn.com/tyk/IjX3R2JMsyaApv0U/img/k8s.png?fit=max&auto=format&n=IjX3R2JMsyaApv0U&q=85&s=055a2af65cf31c5bfcd60aedfd2d8154" width="70" height="69" data-path="img/k8s.png">
    Install on Kubernetes
  </Card>

  <Card href="/5.12/ai-management/ai-studio/installation/linux" img="https://mintcdn.com/tyk/UP13cu4CDQql3M0m/img/linux-icon.svg?fit=max&auto=format&n=UP13cu4CDQql3M0m&q=85&s=cf4da5a1ffd1f401c3bc410c8e9eb3e3" width="64" height="64" data-path="img/linux-icon.svg">
    Install on Linux
  </Card>
</ResponsiveGrid>