Tutorials
This section walks you through building agents with AmritaCore step by step. Each tutorial builds on the previous one and uses the public API only — no internals required.
Prerequisites
- Python 3.10+ with
amrita-coreinstalled (see Getting Started) - An LLM API endpoint (OpenAI-compatible or Anthropic)
- Basic familiarity with
async/await
Tutorial Path
| # | Tutorial | What you will build |
|---|---|---|
| 1 | Create Your First Agent | A minimal chat agent with create_agent() and ChatObject |
| 2 | Add Tools to Your Agent | Register callable tools with @simple_tool and @on_tools |
| 3 | Streaming and Callbacks | Stream responses, read metadata, use callbacks |
| 4 | Events and Hooks | Intercept the pipeline with lifecycle events |
| 5 | Memory and Sessions | Persist conversation history with session_id |
Each tutorial takes about 5–10 minutes. Already comfortable? Jump to Concepts to understand the internals, or Extensions & Integration to plug in your own tools, adapters and MCP servers.