A desktop app for Amp
AgentVera opens the amp command in a real terminal. A new Amp thread is created ahead of time for every agent and the agent always attaches to it. Instead of a model, you pick Amp’s mode.
Amp in AgentVera
- Command
amp- Install
curl -fsSL https://ampcode.com/install.sh | bash- Sign-in
- Amp uses its own ampcode.com login; if you’re signed out, the agent’s terminal asks you to sign in on first launch.
- Model
- Amp uses modes instead of models; the mode decides the model, prompt and tools.
- Effort
- Low, medium, high and ultra modes (--mode).
- Role
- Amp has no extra-instructions flag; use AGENTS.md or rule files for lasting instructions.
- Session resume
- The agent attaches to a thread created with amp threads new and returns to it with threads continue on later launches.
- MCP and browser
- AgentVera’s MCP servers and browser are passed with --mcp-config.
What every agent gets
Side-by-side panes
The agent opens in the same grid as the project’s other agents; the layout is remembered per project.
Flows
When the agent finishes a turn, its reply goes through a template to another agent; turn limits stop loops.
Automatic code review
When a turn ends, the code it wrote is reviewed by the Claude model you pick, with findings linked to file:line.
Telegram notifications
Finished turns, errors and flows waiting for approval reach your phone; reply to message the agent.
Queue and broadcast
Queue a message while the agent is busy, or send one message to the agents you pick.
Picks up where it left off
An agent that has received a message returns to its conversation when the app reopens.
Questions
Can I choose a model in Amp?
Amp has no model flag; you pick the low, medium, high or ultra mode instead.
Does every Amp agent get its own thread?
Yes. A new thread is opened for each agent and the agent returns to it when restarted.
Other agents
All agentsBring your Amp agents into one window.
Download AgentVera, add a project folder and pick Amp under New agent.