Skip to main content
Pylon is a self-hosted daemon you run on your own server. When an event arrives—a Sentry alert, a GitHub pull request, a scheduled cron job—Pylon spins up a sandboxed AI agent in Docker, hands it your codebase, and sends the result back to your chat channel. You stay in control: approve before the agent runs, review the output, and keep your code on your own infrastructure.

Quick Start

Install Pylon and trigger your first AI agent run in minutes.

Installation

Download the binary and set up dependencies for Linux.

What is a Pylon?

Learn the core building block: triggers, agents, channels, and workspaces.

CLI Reference

Full reference for every Pylon command.

​
How Pylon works

1

Install Pylon

Download the binary with a single curl command and run pylon setup to configure your channel and AI agent.
2

Construct a pipeline

Use pylon construct to define what triggers your pipeline (webhook or cron), which AI agent handles it, and where results are sent.
3

Start the daemon

Run pylon start to power up all your pipelines. Pylon listens on a configurable port and waits for events.
4

Receive results in chat

When a trigger fires, Pylon can ask for your approval first, then runs the agent and posts the output directly to your Telegram or Slack channel.

​
Explore by use case

Sentry error triage

Automatically investigate errors and suggest fixes when Sentry fires a webhook.

GitHub PR review

Review pull requests for bugs and security issues on every opened PR.

Scheduled code audit

Run a weekly AI audit of your codebase on a cron schedule.

Custom pipeline

Build a pipeline from scratch with any trigger, agent, and prompt.