Examples

Runnable patterns for agents, teams, workflows, tools, knowledge, and AgentOS.

The Agno cookbook contains examples for building agents, coordinating teams, and running workflows. Start with the basics, then choose a capability below.

Each example lists its setup requirements. Some use local files or databases; others need provider credentials, a running service, or permission to perform an external action. Follow any adaptation notes alongside the cookbook code before running it.

Cookbook links identify the reviewed source revision; adaptation notes explain any changes needed to run that code.

Quickstart

SectionDescription
BasicsYour first agents with storage, memory, knowledge, guardrails, HITL and more.

Primitives

SectionDescription
AgentsAgent patterns: tools, storage, memory, knowledge, learnings, guardrails, multimodal, learning, HITL, hooks, skills and more.
TeamsMulti-agent coordination: modes, structured I/O, human-in-the-loop, knowledge, guardrails, dependencies and more.
WorkflowsWorkflow orchestration: sequential, conditional, parallel, loops, branching, and CEL.

Context

SectionDescription
StorageDatabase backends: PostgreSQL, MongoDB, SQLite, MySQL, Redis.
KnowledgeVector databases, embedders, readers, chunking, filters, and search.
MemoryUser memory persistence across runs, sessions, and agents.
LearningAgents that learn, adapt, and improve from interactions.

Models

SectionDescription
ModelsProvider examples for OpenAI, Anthropic, Google, local models, and more.

Tools

SectionDescription
ToolsMCP, custom tools, search, data, communication, and media integrations.

More

SectionDescription
AgentOSAgentOS: clients, interfaces, databases, middleware, scheduling, and tracing.
ReasoningChain-of-thought reasoning with agents, models, tools, and teams.
EvalsEvaluation patterns: accuracy, agent-as-judge, performance, and reliability.
IntegrationsExternal platforms: observability, RAG integrations, Discord, and more.
ComponentsSave and load agents, teams, and workflows to/from databases.