Guides

Reaching your agent off the terminal

How to get notified — or get a call — when Claude Code and other coding agents finish or get stuck.

How to Get a Phone Call When Claude Code Finishes (or Gets Stuck)

Claude Code goes quiet for hours waiting on one decision. Here's how to get notified when it finishes — or make your phone actually ring when it needs input.

Using the Claude Code Stop Hook to Trigger a Real Notification

Claude Code's Stop and Notification hooks can run any command when your agent stops or waits. Here's how to wire them to a push — or to a phone call.

Claude Code Notifications: Every Way to Get Alerted (Bell, Hooks, Slack, Push, Phone)

A straight comparison of every way to get notified when Claude Code needs you — terminal bell, desktop, ntfy, Pushover, Slack, and a phone call — and when each wins.

How to Get a Phone Call When Codex Finishes (or Needs a Decision)

Running long Codex CLI tasks? Get notified when Codex finishes or pauses for approval — from its terminal notification up to a phone that actually rings.

What Is an MCP Notification Server? (And How to Make Your Agent Reach You)

An MCP notification server lets your coding agent reach you when it's blocked. Here's what that means in MCP terms, and how a tool-based one rings your phone.

The Best MCP Servers for Long-Running Coding Agents

The MCP servers a long unattended coding run actually needs — memory, filesystem, search, and the human-escalation slot most setups forget. A category map.

AI Agents That Call Your Phone: The Case for Voice Escalation

Why a phone call beats a chat ping when an AI coding agent needs you: a ring demands the present tense, voice resolves more, and your reply is the result.

Human-in-the-Loop Coding Agents: When Should an Agent Stop and Ask?

A practical framework for when an autonomous coding agent should stop and ask a human — irreversible actions, ambiguity, approvals, spend, and security.

Confirming Destructive Actions by Voice: Why Mistranscription-Safe Approvals Matter

Speech-to-text can mishear a destructive command. Here's how AgentCall's voice approval confirms irreversible actions out loud before an AI agent runs them.

Running Coding Agents Unattended Overnight Without Babysitting

Leaving Claude Code or Codex running overnight, then finding it stalled on one question? Make unattended runs safe, gated, and reachable by a phone that rings.

How to Monitor a Long-Running AI Coding Agent Without Watching the Terminal

Kick off a multi-hour agent run and walk away safely. How to monitor a long-running coding agent — logs, hooks, push, and a phone call — and where each fits.

Make Any MCP Agent Call You When It Needs a Decision

AgentCall is an MCP server that calls you — any MCP agent that allows long-running tools (Claude Code tested, Codex beta) can ring your phone for a decision.