agents/examples/python_agent_nodes/hello_world_rag

Hello World RAG

Minimal Retrieval-Augmented agent that demonstrates the new Hanzo Agents memory vector APIs without any external services.

What It Does

Component	Purpose
ingest_document skill	Accepts raw text or a file path, chunks it, creates toy embeddings, and stores them via app.memory.set_vector.
answer_question reasoner	Embeds the question, calls app.memory.similarity_search, and feeds the retrieved context into the LLM to synthesize an answer.

Running

cd examples/python_agent_nodes/hello_world
pip install -r requirements.txt
python main.py

Once the agent is registered with the control plane, you can call it via the Hanzo Agents CLI or direct HTTP:

# Ingest a document (either provide text or a path)
hanzo-agents call hello-world-rag ingest_document \
  --json '{"document_id": "guide", "path": "docs/README.md"}'

# Ask a question
hanzo-agents call hello-world-rag answer_question \
  --json '{"question": "What does the README say about deployment?"}'

The output includes both the synthesized answer and the supporting chunks so you can trace where the knowledge came from.