Skip to content

Tutorials

Hand-held end-to-end walkthroughs. Goal: get you to a working result so you trust the library and know where things live, and get you looking at what the model is actually doing rather than only at its final answer.

Work through them in order; each builds on the previous. Total time end-to-end: about 60 minutes.

  1. Getting started (~15 min): install AIMU, send your first message, build your first agent.
  2. First agent with tools (~10 min): @tool decorator and the agentic loop.
  3. Workflows (~15 min): Chain.from_client, Router.from_client, Parallel.from_client for code-controlled patterns.
  4. Streaming (~10 min): the StreamChunk API across chats, agents, and workflows.
  5. Vision & audio input (~10 min): send images and audio to a model.

Once you've finished, the how-to guides cover specific tasks; the explanation pages cover the why behind the design.

What you'll need

  • Python 3.11+
  • One model backend. The tutorials use Ollama because it's local and free, but every snippet works with any provider; just swap the model string:
aimu.chat("hi", model="ollama:qwen3.5:9b")        # local
aimu.chat("hi", model="anthropic:claude-sonnet-4-6")  # cloud
aimu.chat("hi", model="openai:gpt-4o-mini")          # cloud

For Ollama: install from ollama.com and run ollama pull qwen3.5:9b once. For cloud providers, set ANTHROPIC_API_KEY / OPENAI_API_KEY / GOOGLE_API_KEY in your environment or a .env file.