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

# Memory Management

Memory management is essential for building intelligent agents that can maintain context over multiple conversational turns (multi-turn dialog). Lyzr Studio utilizes **Lyzr Cognis**, a production-grade memory module, alongside flexible options for integrating external cloud infrastructure.

***

## 🚀 Introducing Lyzr Cognis (Default Provider)

**Lyzr Cognis** is the default memory provider for all agents created in Lyzr Studio. It is a state-of-the-art module built specifically for high-stakes, production-grade workloads where understanding the "long-tail" of a conversation is critical.

### **Key Features of Cognis:**

* **Temporal Awareness:** Understands the timing and sequence of interactions for superior historical context.
* **Cross-Session Memory:** Unlike standard short-term memory, Cognis can remember details about a specific user across different sessions. If a user returns days later, the agent retains knowledge of previous interactions, preferences, and outcomes.
* **Matryoshka Embeddings:** Utilizes advanced embedding techniques to optimize search accuracy and storage efficiency.
* **SOTA Benchmarking:** Competes against and outperforms industry leaders across several retrieval and accuracy parameters.

***

## ⚙️ Configuring Short-Term Memory

Short-term memory stores recent messages from the current session, ensuring the agent remembers immediate context without users having to repeat themselves.

### **Step 1: Enable Memory and Select Provider**

1. **Locate:** In the **Create Agent** view, find the **Memory** toggle under **Core Features**. Toggle it **ON**.
2. **Configure:** Click the settings icon next to the Memory feature to open the **Configure Memory** modal.
3. **Select Memory Provider:** Choose the desired provider from the updated dropdown menu:
   * **Lyzr Cognis (Default):** The recommended high-performance provider featuring temporal and cross-session awareness.
   * **Lyzr Memory:** A balanced, built-in memory storage option for simpler use cases.
   * **Amazon Bedrock AgentCore Memory (BYOM):** Lyzr supports **Bring Your Own Memory (BYOM)**, allowing you to leverage your own AWS infrastructure.

<img src="https://mintlify.s3.us-west-1.amazonaws.com/lyzrinc/assets/images/studio/mem1.png" alt="Configure Memory Interface" />

***

## ☁️ AWS External Memory Configuration (BYOM)

For organizations requiring maximum data control, selecting the **Amazon Bedrock AgentCore Memory** option allows you to integrate your own cloud environment.

* **Cloud Integration:** Use your own AWS account to store and manage agent memory data.
* **Credentials:** You must provide the necessary **AWS Credentials** to enable the BYOM connection.
* **Resource Selection:** Specify the existing memory resource within your AWS console to house the agent's conversational data.

<img src="https://mintlify.s3.us-west-1.amazonaws.com/lyzrinc/assets/images/studio/memory3.png" alt="" />

***

## 🔄 Memory Provider Comparison

| Provider           | Type         | Core Strengths                                                                  |
| :----------------- | :----------- | :------------------------------------------------------------------------------ |
| **Lyzr Cognis**    | **Advanced** | **Cross-session persistence**, temporal context, and production-scale accuracy. |
| **Lyzr Memory**    | **Standard** | Balanced performance for simple, single-session interactions.                   |
| **Amazon Bedrock** | **BYOM**     | Enterprise-grade privacy; data stays within your private AWS VPC.               |

***

## ✅ Summary: Memory Behavior

* **Coherence:** The agent uses recent messages to maintain fluid, natural dialogue.
* **User Persistence:** With **Cognis**, memory is no longer session-scoped; the agent builds a profile of the user that persists across multiple calls or chats.
* **Flexible Storage:** Choose between Lyzr's optimized SOTA modules or your own enterprise cloud storage for complete data sovereignty.