Agent Tick Docs
Agent Tick is a least-permission approval layer for coding agents. Your local agent can send progress, ask a bounded question, or wait for approval without giving the phone, hosted app, or Agent Tick server a shell.
Use Agent Tick for three things:
- Status Updates — one-way progress from an agent.
- Steering — choose from bounded options supplied by the agent.
- Sanctions — approve or deny one specific risky action before the agent continues locally.
In Shared Workspaces, routed Requests can also become Team Decisions. Broader bounded Steering polls use Community Votes through Audience Channels.
Agent Tick routes requests and returns bounded responses. It does not execute commands remotely.
Start here
- Quick Start — connect the app, send a Test Request, then connect your first Agent Connection.
- Core Concepts — understand Requests, Sessions, Agent Connections, Approval Devices, and Workspaces.
- Team Decisions and Community Votes — understand Shared Workspace decisions and Audience Channel Requests.
- Prompt agents to use Agent Tick — make agents call the Agent Tick tool instead of just replying in chat.
Common paths
| I want to… | Read |
|---|---|
| Connect my first machine | Quick Start |
| Understand the app screens | Native App and Personal Console |
| Understand Private Requests and private Status Updates | Private encryption |
| Configure Claude Code, Codex, Pi, or other tools | Coding-agent integrations |
| Try Agent Tick from a terminal | CLI |
| Use Agent Tick with a team | Workspaces and Team Decisions |
| Ask a broader audience to vote | Community Votes |
| Run my own server | Self-hosting Quick Start |
| Build a custom integration | API and SDK |
Privacy Policy and Terms live on the product site: