Local project preview

Support Swarm

A self-improving agent team for triaging customer requests and safely completing routine actions.

Open GitHub repository Visit project website Watch demo video

Project description

Support Swarm coordinates focused agents, scores the quality of each completed run, and uses reviewer feedback to improve the next attempt. The team built the full workflow during the hackathon, using Pioneer for the core agent loop and Guild.ai for evaluation, communication, or runtime support. The demo includes a visible activity trace so judges can inspect what each agent did and where human approval was required.

Project links

Technical architecture

An OpenAI Agents SDK router classifies each request, hands it to a constrained specialist, and requires human approval before any account-changing tool call.

Setup instructions

Clone the public repository, install dependencies, copy .env.example to .env.local, add the documented API keys, and run the development command in the README.

Lessons learned

The team found that clear handoffs and fast evaluation loops mattered more than adding more agents. Pioneer handled orchestration while Guild.ai made the result easier to inspect and improve.

Additional screenshots or links

https://openai.github.io/openai-agents-python/
https://github.com/openai/openai-agents-python

Tools used

Pioneer

Guild.ai

Back to project gallery