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

# Prerequisites

> Review the prerequisites for installing Cognigy Agent Copilot, including Cognigy, Kubernetes namespace, and Helm chart requirements.

> [!NOTE]
>
> If you are a new user of Agent Copilot and are starting to use it from [Cognigy version 4.65](https://docs.cognigy.com/release-notes/4.65.md), you don't need this installation guide. From 4.65 onwards, Agent Copilot is included with the [Cognigy installation](https://docs.cognigy.com/ai/installation/overview/).

Before installing and running Agent Copilot, you need to have Cognigy installed and a dedicated `agent-assist` namespace created in your Kubernetes cluster. Helm, a package manager for Kubernetes, is also required to streamline the deployment of the necessary resources.

* [Kubernetes](https://kubernetes.io/) 1.16+
* [Helm](https://helm.sh/) 3.1.0+

## Replication and Resource Allocation

The default number of replicas is set to 3 in order to maintain high availability standards. The Agent Copilot pods are stateless, allowing you to scale them up or down based on demand. The agent-assist-backend pod bears the brunt of the workload and should be the primary candidate for scaling to accommodate an increase in concurrent users.

### MongoDB

For detailed MongoDB setup and configuration, refer to the [Cognigy Installation Prerequisites guide](https://docs.cognigy.com/ai/administer/installation/prerequisites#kubernetes-cluster).

### Redis

Redis is used to store WebSocket subscriptions and cached Agent Copilot session parameters, so its storage requirements are minimal. You can start with a 1GB allocation and scale up as necessary.