Using Teams for Multi-Agent Collaboration
Teams let you group multiple AI agents together. When you send a message to a team, a supervisor agent analyzes your request and routes it to the most appropriate team member.
Why use teams?
Instead of manually picking the right agent for each question, teams handle routing automatically. A team with a Code Analyst, Senior Developer, and Software Architect can handle any programming question — the supervisor figures out which specialist is best suited.
Creating a team
- Navigate to Teams in the sidebar.
- Click Create Team and give it a name and description.
- Add agents to the team — pick agents with complementary skills.
- The supervisor agent is assigned automatically.
How routing works
When you send a message, the supervisor:
- Analyzes your request
- Selects the best team member based on their specialties
- Routes your message to that agent
- Returns the response with the responding agent identified
You can also @-mention a specific team member to bypass routing — type @agent-name in your message to direct it to a particular specialist.
Team visibility
Teams support visibility levels:
- Public — Anyone on the platform can chat with your team.
- Organization — Only members of your organization can access it.
- Private — Only you can use it.
Assigning teams to projects
You can bind a team to a project (up to 5 teams per project). When a team is assigned to a project, project members can use that team in their conversations and benefit from the project's shared knowledge files and system prompt.
Tips for effective teams
- Keep teams focused — a "Full-Stack Development" team with 3-5 coding agents works better than a 10-agent team that covers everything.
- Include agents with complementary skills, not overlapping ones.
- Name and describe your team clearly so the supervisor understands the team's purpose.