In brief

OCI GenAI Python Starters is a collection of small runnable Python demos for Oracle Cloud Infrastructure Generative AI. It helps developers move from credentials to a working example using the OCI SDK, LangChain, LlamaIndex, LangGraph, or OpenAI-compatible API flows.

Every time I want to try something new on OCI Generative AI, I end up needing the same thing: one tiny Python example that actually works.

Not a full framework. Not a polished app. Just a small script with the right config, the right endpoint, and one clear path from “I have credentials” to “okay, it runs.”

After rebuilding those snippets a few times, I put them in one repo: OCI GenAI Python Starters .

Nothing fancy, just a bunch of small demos you can run, break, tweak, and reuse.

What’s in there

  • direct-sdk if you want to start from the bare minimum
  • langchain, llamaindex, and langgraph if you want to see the same idea through higher-level tools
  • openai-chat, openai-agents, and openai-responses if you prefer the OpenAI-style route
  • One folder per demo, with its own requirements.txt and .env.example
  • A test_all_demos.sh script if you want to run through everything once the environment variables are in place

I also tried to keep the order sensible: start from the raw SDK, then move to the framework examples once the basic OCI flow is clear.

One thing that tripped me up

The OpenAI-compatible examples are not all wired the same way.

openai-chat and openai-agents still use the regular OCI Generative AI endpoint. openai-responses is the odd one out: it uses the OpenAI-compatible base URL and also needs a Generative AI project OCID.

It is a small difference, but it is exactly the kind of detail that can waste half an hour when all you wanted was a quick hello-world.

Anyway, if you’re playing with OCI GenAI in Python and want a few simple starting points, the repo is here: OCI GenAI Python Starters .

FAQ

What is OCI GenAI Python Starters?

OCI GenAI Python Starters is a collection of small runnable Python examples for Oracle Cloud Infrastructure Generative AI, including direct SDK usage, LangChain, LlamaIndex, LangGraph, and OpenAI-compatible API flows.

Who should use OCI GenAI Python Starters?

Cloud engineers, developers, and solution architects can use it when they need a quick working starting point for OCI Generative AI experiments in Python.

Which OpenAI-compatible OCI example needs a Generative AI project OCID?

The openai-responses example uses the OpenAI-compatible base URL and needs a Generative AI project OCID.