# Team view

URL: https://agentvera.dev/en/features/team-view/ · Plan: Free

Instead of steering agents one by one, build a team. The agents open in your workspace go onto an org chart: one becomes the manager, the others get a role, instructions and someone to report to. You hand the work to the team; the manager splits it, hands it out and checks the results. The team view then shows who is working, who is messaging whom and where each task stands, on one screen.

**Team structure: an org chart of agents.** The Team structure window opens from the team icon on a group or at the top of the sidebar. The agents of your workspace wait in a pool; click or drag them onto the chart. A team holds up to 24 agents and one manager.

- Each agent has a role, instructions, someone it reports to and someone it hands its work to.
- Drop a card on another to hand it the work, report to it or swap places.
- Arrange automatically tidies the chart; Open a new agent opens an agent with any installed CLI and adds it to the team.
- Saving puts the agents into the group; Save and open the team goes straight to the team view.

**Ready-made roles and teams.** Ready-made roles fill in the name and instructions of a researcher, planner, developer, reviewer, tester, DevOps, designer, documentation or security agent; you can change all of them, and several agents can share a role. Ready-made teams open the agents they need for you.

- Software team: tech lead, planner, developer, reviewer and tester.
- Research team: research lead, researcher and report writer.
- Social media team: a manager, trend researcher, copywriter, visual designer and ads specialist; the ads specialist drafts and asks before spending.

**MCP servers for the team and its roles.** Give the team MCP servers and every agent in it starts with them. A role’s own servers come on top, so the tester works with Playwright and the designer with fal.ai and Canva. An agent can opt out of the team’s servers in its own settings. Changed agents restart in the same conversation once their turn is over. If a ready-made team suggests a server you have not added yet, you are told and can add it from the MCP catalog.

**Hand the work to the team.** Nothing starts when the team is built. You give the work in the box below; it goes to the manager by default, or to any agent you pick. The agents are first briefed on their roles, and briefed again only when the team changes. Agent chat is turned on if it was off; the manager splits the work, hands it out and checks the results.

**The live team view.** The group’s Team tab shows the team in place of its terminals. Working agents glow with a calm ring, messages travel along the lines as a light, and a speech bubble appears over the sender. The attention bar on top lets you approve or refuse permission requests in place, and shows an agent that failed and the manager’s summary when it is done.

- Team status panel: what each agent is working on, its current step (editing, reading, running, searching, planning, messaging) and the time spent.
- The feed tab filters messages, work and agent events; the tasks tab shows the shared tasks and their progress.
- Status report asks the manager for a short report; Interrupt all stops every working agent.

**A closer look at one agent.** Click a card to open the agent panel: its last reply, messages, tasks, recent steps, the files it changed and its live terminal. Choose Interrupt and redirect for a working agent, or Queue your message for the end of its turn. Switching to another group brings back that group’s terminals.

**Build your first team:**

1. Open Team structure from the team icon at the top of the sidebar or on a group.
2. Pick a ready-made team, or drag agents from the pool onto the chart and make one the manager.
3. Give each agent a role and instructions; draw whom it reports to and whom it hands its work to.
4. Add MCP servers for the team and its roles if you like, then choose Save and open the team.
5. Type the work in the box below and choose Give to the team; the manager splits it while you watch the team view.

- **Which agents can be in a team?** Any agent open in your workspace: Claude Code, Codex or any other installed CLI agent. Open a new agent adds one without leaving the window. A team holds up to 24 agents.
- **Do the agents start working as soon as the team is built?** No. The team waits until you give it work in the box below. Then the agents are briefed on their roles and the manager hands out the work.
- **How do the agents give each other work?** Through agent chat, which is turned on if it was off. Agents send messages, wait for replies and use the shared task list. Every message shows up in the team view as a light travelling along a line.
- **What happens when an agent asks for permission?** The request shows up in the attention bar above the chart, and you can approve or refuse it without going to the terminal.
- **How is the team view different from a parallel team?** The team view turns your open agents into a lasting team that works together while you watch. A parallel team is an agent splitting one job into parts, running each part in its own worktree and merging them into one branch.

---

**About AgentVera.** AgentVera is a desktop app for macOS, Windows and Linux, built by Solviera Teknoloji. It runs 18 AI coding agents (Claude Code, Codex, Gemini CLI, Copilot, Cursor, Antigravity, opencode, Hermes, Qwen Code, Kimi Code, Kiro, Amp, Droid, Goose, Vibe, Crush, Aider, Cline) side by side, per project, in real terminals with the user's own CLI installs and logins. The free plan has unlimited agents. Download: https://agentvera.dev/en/download/ · Pricing: https://agentvera.dev/en/pricing/