I’ve been experimenting with AI tools for a while and kept running into the same problem: they’re great for individuals, but painful for teams. Everything is scattered — prompts in chats, workflows in separate apps, agents running on different services. Sharing or organizing work with others was always clunky.

So I started building Weam, an open-source project that tries to make AI more “team-friendly.”

What it does today:

Organize prompts, chats, and agents into Brains (team folders).

Run both lightweight agents and more complex Pro Agents.

Plug in your own LLM keys (OpenAI, Anthropic, Gemini, Llama, etc.).

Set up RAG pipelines to use your own documents.

Self-host, so teams keep control of their data.

Repo: https://github.com/weam-ai/weam

Docs / site: https://weam.ai/

This is the first public release that feels usable. I’d love feedback on:

What’s missing for real team adoption?

Whether the “Brains” model makes sense for collaboration.

Any technical blockers you see in trying it out.

HN thread from the earlier intro (not a full Show HN): [link if you want to include the old one].

Would really appreciate thoughts from anyone who’s tried to get AI working in a team environment.