Skip to content

FAQ

Common questions about Bee. For questions about AI engineering itself, check the relevant section or ask in Discussions.

About the project

Is Bee affiliated with any AI company?

No. Bee is an independent, community-driven, open-source project. Examples use vendor SDKs, but the concepts are vendor-neutral.

Why 'Bee'?

A hive is a community that builds something bigger than any individual โ€” exactly what this repository is. Plus, knowledge is the honey. ๐Ÿฏ

Is Bee free?

Yes. Content is CC-BY-4.0 and code is MIT. Use it in your courses, blogs, and projects โ€” just give credit.

Using the content

Which LLM provider do the examples use?

Concepts are provider-agnostic. Code defaults to the Anthropic SDK for concreteness, with notes on adapting to OpenAI, Google, Ollama, and others. Most examples run against any provider with minor changes.

Do I need a GPU?

No. Almost everything uses hosted APIs. The few local-model sections say so explicitly and offer CPU-friendly alternatives.

Will the code examples go stale as APIs change?

We fight this two ways: each example pins its dependencies, and CI runs example tests. When a provider makes a breaking change, it surfaces as a failing check and gets fixed. Found one that's broken? Open an issue.

How much will it cost me to run the examples?

Cents, usually. Examples use small models and short prompts by default, and always set max_tokens. Each example's README notes anything unusual.

Learning with Bee

I'm a complete beginner. Where do I start?

Getting Started, then the AI Engineer โ€” Fundamentals path. Don't read randomly โ€” follow a path.

I'm experienced. Can I skip around?

Absolutely. Jump straight to RAG, Agents, or any topic. Difficulty tags (๐ŸŸข๐ŸŸก๐Ÿ”ด) tell you what a page assumes.

How is this different from a framework's docs?

Framework docs assume you already know the concepts and want you to use their tool. Bee teaches the concepts first and stays vendor-neutral, so the understanding transfers no matter what you build with.

Contributing

I found a mistake. What do I do?

Open an issue โ€” or fix it and open a PR! See Contributing.

Can I add content about a topic that isn't here yet?

Yes, please! Check for a [WANTED] marker in the relevant section, claim it, and use the templates. See CONTRIBUTING.md.

How do I get credited?

Every merged contribution earns you a spot via the All Contributors spec โ€” docs, code, design, and ideas all count.


Didn't find your answer?

Ask in Discussions or read SUPPORT.md.