Guides

Step-by-step guides for building with PraisonAI

Learn how to build AI agents with PraisonAI through practical, step-by-step guides.

from praisonaiagents import Agent

agent = Agent(
    name="Assistant",
    instructions="You are a helpful assistant.",
)

agent.start("Hello! What can you help me with?")

The user starts with a guide that matches their goal, runs the sample agent, then extends it step by step.

Single Agent

Build your first AI agent

Multi-Agent Systems

Orchestrate multiple agents working together

Workflows

Create complex agent workflows

RAG & Knowledge

Add knowledge bases to your agents

Persistence

Save and resume agent sessions

Deployment

Deploy agents to production

Quick Start

1. Install PraisonAI

   pip install praisonaiagents
   export OPENAI_API_KEY=your_api_key
   
2. Create your first agent

   from praisonaiagents import Agent
   
   agent = Agent(
       name="Assistant",
       instructions="You are a helpful assistant.",
   )
   
   agent.start("Hello! What can you help me with?")
   

By Topic

Agent Patterns

• Single Agent - Build a single AI agent
• Multi-Agent Systems - Multiple agents working together

Workflows

• Routing Workflow - Route tasks to specialized agents
• Parallel Workflow - Run agents in parallel
• Sequential Workflow - Chain agents sequentially
• Orchestrator Pattern - Central orchestrator managing agents

RAG & Knowledge

• Knowledge Base Setup - Create and configure knowledge bases
• Chunking Strategies - Optimize document chunking
• Retrieval Methods - Configure retrieval strategies

Persistence

• Overview - Persistence concepts
• Database Setup - Configure database backends
• Session Resume - Resume interrupted sessions

Deployment

• Overview - Deployment options
• Docker - Deploy with Docker

Best Practices

Start with a single agent

Begin with one agent before building multi-agent systems. Get comfortable with instructions and tools first.

Use clear, specific instructions

Agent instructions should be specific and action-oriented. Vague instructions lead to inconsistent behavior.

Add tools incrementally

Add one tool at a time and verify each works before combining multiple tools in one agent.

Test with simple tasks first

Validate your agent with simple tasks before giving it complex, multi-step workflows.