Skip to main content
pylon setup walks you through an interactive wizard that creates your global configuration. Run it once before creating any pipelines. It checks that Docker is installed, prompts for your notification channel and AI agent preferences, and writes credentials to ~/.pylon/.env.

​
Usage

pylon setup

​
What setup configures

1

Docker check

Setup verifies Docker is installed and the daemon is running. If Docker is not found, it prints the install URL and continues.
2

Default channel

Choose where Pylon sends notifications and approval requests:
  • Telegram — requires a bot token from @BotFather and a group chat ID
  • Slack — requires a bot token, app-level token, and channel ID
  • stdout — prints to the console only (useful for testing)
  • Webhook — generic HTTP POST (no built-in UI; configure endpoint manually)
Discord, WhatsApp, and iMessage support is coming soon.
3

Default AI agent

Choose the AI agent that runs inside Docker containers:
  • Claude Code — authenticate via OAuth (your existing ~/.claude/ session) or an ANTHROPIC_API_KEY
  • OpenCode — use built-in Zen (no key needed) or bring your own API key (Anthropic, OpenAI, or Google)
4

Public URL (optional)

Enter the base URL where external services can reach Pylon (e.g. https://my-server.example.com). Leave blank if you only run locally or will configure this later.
5

Agent image build

Setup builds (or verifies) the Docker image for your chosen agent. This may take a few minutes on the first run.

​
Telegram setup

When you select Telegram, setup guides you through:
  1. Creating a bot via @BotFather (/newbot)
  2. Copying the bot token
  3. Detecting your group chat ID automatically (add the bot, send a /command, Pylon detects the ID) or entering it manually
The bot token is saved to ~/.pylon/.env as TELEGRAM_BOT_TOKEN. The config references it as ${TELEGRAM_BOT_TOKEN}.

​
Slack setup

When you select Slack, setup displays this app manifest to paste into the Slack API console: 
display_information:
  name: Pylon
  description: AI agent pipeline runner
features:
  bot_user:
    display_name: Pylon
    always_online: true
oauth_config:
  scopes:
    bot:
      - chat:write
      - channels:history
      - channels:read
      - groups:history
      - groups:read
      - reactions:write
settings:
  event_subscriptions:
    bot_events:
      - message.channels
      - message.groups
  interactivity:
    is_enabled: true
  socket_mode_enabled: true
Create the app at api.slack.com/apps, enable Socket Mode, and paste in your bot token (xoxb-...) and app-level token (xapp-...).

​
Output files

FileContents
~/.pylon/config.yamlChannel type, agent type, server settings
~/.pylon/.envSecrets (bot tokens, API keys) — loaded automatically on pylon start

​
Next steps

After setup, create your first pipeline: 
pylon construct my-sentry --from sentry
pylon start