DUMB DEV Community

Vishal Kachalia
Vishal Kachalia

Posted on

We Open-Sourced Our AI Platform - Here's What We Learned

After building WEAM internally for 2+ years, we made the decision to open source the entire platform. Here's the story and what we're hoping for from the community.

What is WEAM?
WEAM is essentially "ChatGPT for teams" - but with some pretty interesting additions:

  • Multi-LLM support: OpenAI, Claude, Gemini, Llama, and more
  • Organizational structure: "Brains" that organize AI interactions by team/project
  • Custom agents: Build specialized AI assistants with knowledge bases
  • RAG pipeline: Upload documents and have AI reference them intelligently
  • Ready-made apps: QA analysis, video processing, content generation, etc.

Tech Stack

  • Frontend: Next.js (fully responsive)
  • Backend: Node.js + Python
  • Deployment: Docker-ready, 2-minute setup
  • Integrations: MCP protocol, OAuth 2.0, JWT auth

Why Open Source?
Honestly? Three reasons:

  • AI platforms shouldn't be black boxes 📦
  • The community will probably build cooler stuff than we ever could 🚀
  • We're genuinely curious what happens when developers get their hands on it 😄

What We're Looking For

  • Code reviews (especially architecture feedback)
  • Bug reports (they're definitely hiding in there)
  • Feature requests for our roadmap
  • Contributors who want to build cool AI apps with us

Get Started
bashgit clone https://github.com/weam-ai/weam.git
cd weam
bash build.sh
docker-compose up -d

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

What do you think? Ready to help us figure out what we got right (or hilariously wrong)?

Top comments (0)