> ## 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.

# Cognis Memory Cookbooks

> Integration guides for adding persistent memory to your AI agents

Step-by-step guides for integrating Cognis memory with popular agent frameworks. Each cookbook includes working code for both the **hosted** (lyzr-adk) and **open-source** (lyzr-cognis) versions.

## Getting Started

<CardGroup cols={2}>
  <Card title="Your First Memory" icon="rocket" href="/cognis/cookbooks/getting-started">
    Add, search, and retrieve memories in 3 minutes. Covers both hosted and open-source.
  </Card>

  <Card title="Hosted vs Open Source" icon="code-compare" href="/cognis/comparison">
    Not sure which version to use? Compare features, APIs, and trade-offs.
  </Card>
</CardGroup>

## Framework Integrations

<CardGroup cols={2}>
  <Card title="Cognis + CrewAI" icon="users" href="/cognis/cookbooks/cognis-crewai">
    Inject user preferences into multi-agent crew backstories and store results for future runs.
  </Card>

  <Card title="Cognis + Agno" icon="robot" href="/cognis/cookbooks/cognis-agno">
    Expose memory as tools that agents autonomously decide when to use.
  </Card>

  <Card title="Cognis + LangGraph" icon="diagram-project" href="/cognis/cookbooks/cognis-langgraph">
    Memory as graph nodes in a retrieve → respond → store pipeline.
  </Card>

  <Card title="Cognis + LangChain" icon="link" href="/cognis/cookbooks/cognis-langchain">
    LCEL chain integration with memory injection via MessagesPlaceholder.
  </Card>
</CardGroup>

## Integration Patterns

| Framework | Pattern          | Description                                                       |
| --------- | ---------------- | ----------------------------------------------------------------- |
| CrewAI    | Manual injection | Retrieve → inject into backstory → run crew → store results       |
| Agno      | Tool-based       | Memory ops as `@tool` decorators; agent decides when to use       |
| LangGraph | Graph nodes      | Dedicated nodes for retrieve/store in explicit control flow       |
| LangChain | LCEL chains      | Memory retrieval as prompt injection; store after chain execution |