Quickstart

Install the CLI

One command on macOS or Linux:

curl -fsSL https://polylane.com/install | bash

The script downloads the latest release of the polylane CLI, verifies its checksum, installs it to ~/.polylane/bin, and puts it on your PATH. It then runs polylane setup, which detects your coding agents (Claude Code, Cursor, OpenCode, Codex, Windsurf, Zed, VS Code), installs the Polylane skill where the agent supports skills, and registers the platform MCP server in each agent's own config — skip that step with --no-setup. It needs Node.js 20+ and falls back to Homebrew when Node is missing. Safe to re-run. Prefer a package manager, or on Windows? Every channel is listed in the CLI overview.

Then sign in:

polylane auth login

This opens the browser OAuth flow (use --no-browser on SSH or headless machines). No account yet? Create your workspace at console.polylane.com first. From here the whole platform is in reach: polylane issue list --active shows what needs attention, and every command group is a --help away.

What is Polylane?

Polylane runs agents that read your code, watch your infrastructure, detect issues, investigate them, and fix them. The platform works as a loop with four steps:

• Context: connect your clouds, repositories, and observability providers. Polylane builds a live context graph of everything, rendered as your topology.
• Detection: checks run against connected resources and raise issues. Change intelligence records every infrastructure shift. Alerts from your providers flow into the same lifecycle.
• Investigation: agents dig into what was detected through threads and investigations.
• Remediation: autofixes, automations, and skills turn findings into fixes.

Set up with your coding agent

Prefer to let the agent you already use do the setup? Copy this prompt into your coding agent, then follow its lead:

The agent points you at Google or GitHub sign-in, which arrives already verified, then connects the platform MCP server with an API key and connects every cloud and integration it can from credentials you already have. If you prefer email and password, it runs signup itself and asks you for the emailed code. Either way the GitHub App installation and OAuth for Vercel, Slack, and Sentry stay yours to click. Nothing to install.

Or set up from the console

1. Sign in at console.polylane.com and create a workspace.
2. Connect one of each: a cloud, a code host, and an observability provider. See Integrations and Clouds for everything you can connect.
3. Start a thread and ask about what Polylane already sees: "which services run in our production cloud account, and which had deploys in the last 24 hours?"

Watch the loop start

Detection begins as soon as resources sync. Checks run against everything you connected, and what they find lands in Issues. Provider alerts you already have flow into the same list, and every sync records what changed in change intelligence.

Issues and alerts get investigated on their own; you don't need an automation for that. Automations are for the signals only you know about. For example, your support channel is often the first place a real problem shows up:

Use Polylane to set up an automation: on every message in the #support channel in Slack, have an agent triage whether it describes a real problem in production, and open an issue with the findings when it does.

Next steps

• Coding agents: everything your agent can do with Polylane, plus the CLI and the browser extension.
• Concepts: the platform's primitives in one place.
• Context: clouds, repositories, topology, and memories.
• Detection: issues, change intelligence, advisories, and scans.
• Investigation: threads, investigations, and escalations.
• Remediation: autofix, automations, and skills.
• API reference for the full HTTP surface.