Process Manager for Autonomous AI Agents

📰 Hacker News (AI)

botctl is a process manager for autonomous AI agents, allowing users to manage persistent AI bots with a terminal dashboard, web UI, and declarative configuration

intermediate Published 9 Apr 2026
Action Steps
  1. Install botctl using the provided installation script
  2. Create a bot using the interactive creation process or with flags
  3. Launch the dashboard to monitor and control bots
  4. Control bots using the CLI or dashboard, including starting, stopping, messaging, and resuming
Who Needs to Know This

DevOps and software engineering teams can benefit from botctl as it provides a scalable and efficient way to manage AI agents, while data scientists and AI researchers can use it to deploy and monitor their models

Key Insight

💡 botctl provides a declarative configuration and autonomous execution for AI agents, making it easier to manage and deploy AI models

Share This
🤖 Manage autonomous AI agents with ease using botctl! #AI #DevOps

Key Takeaways

botctl is a process manager for autonomous AI agents, allowing users to manage persistent AI bots with a terminal dashboard, web UI, and declarative configuration

Full Article

Published Time: Mon, 06 Apr 2026 14:59:59 GMT

# botctl — Process Manager for Autonomous AI Agents

Copied

[$ botctl](https://botctl.dev/)
* [How It Works](https://botctl.dev/#architecture)
* [Quickstart](https://botctl.dev/#quickstart)
* [Docs](https://botctl.dev/docs)
* [GitHub](https://github.com/montanaflynn/botctl)

[ process manager for ai agents ]

Manage persistent AI bots with a terminal dashboard, web UI, and declarative configuration.

macOS / Linux Windows

$curl -fsSL https://botctl.dev/install.sh | sh

>irm https://botctl.dev/install.ps1 | iex

[GitHub](https://github.com/montanaflynn/botctl)[Quickstart](https://botctl.dev/#quickstart)

botctl

BOTCTL bots: 2/3 active cost: $4.51

NAME STATUS RUNS COST

<logs>weather-bot

[run #12] 2026-02-12 11:42:03

<bash> curl -s api.weather.gov/alerts

✓Fetched 3 active alerts

<write> workspace/alerts.json

No severe storms detected.

Sleeping for 300s...

msg> check the error logs

s stop m message q quit

$

How It Works

Declarative config

01

YAML frontmatter for settings, markdown body for the prompt.

---name: code-reviewer interval_seconds: 60 max_turns: 20--- Review open PRs and post comments...

Autonomous execution

02

Spawns Claude with your prompt, tools, and workspace. Runs, logs, and sleeps on a loop.

$ botctl start code-reviewer -d✓ Harness started (pid 48201) $ botctl logs code-reviewer -f[run #1] Reviewing PR #49...

Session memory

03

Every run saves its session. Resume where Claude left off, or send messages to redirect a running bot.

$ botctl start review --message"focus on PR 51"✓ Message queued, waking bot...

Hot reload

04

Edit `BOT.md` and the next run picks up changes. No restarts, no deploys, no downtime.

# change max_turns in BOT.md max_turns: 20→max_turns: 50# next run uses new value automatically

Extensible skills

05

Search, install, and share reusable skill modules from GitHub. Skills inject capabilities into any bot's prompt.

$ botctl skills search slack $ botctl skills add owner/repo --skill slack-notify $ botctl skills list

Web dashboard

06

Monitor and control bots from the browser. Same capabilities as the TUI — start, stop, message, and stream logs.

$ botctl --web-ui✓ Dashboard at http://localhost:4444 $ botctl --web-ui --port 8080

Quickstart

01

### Install

Download the latest binary with checksum verification. Supports macOS, Linux, and Windows on AMD64 and ARM64.

macOS / Linux Windows

$curl -fsSL https://botctl.dev/install.sh | sh

>irm https://botctl.dev/install.ps1 | iex

02

### Create a bot

Interactive creation via Claude. Generates a BOT.md config file with your bot's name, schedule, and system prompt.

shell

$botctl create my-bot# Opens interactive creation with Claude# Or with flags:$botctl create my-bot-d"Monitor weather APIs"-i 300-m 20

03

### Launch the dashboard

Open the TUI dashboard to monitor and control all your bots. Or use the web UI for browser-based access.

shell

# Terminal dashboard$botctl# Web dashboard$botctl--web-ui# port 4444$botctl--web-ui --port 8080

04

### Control your bots

Start, stop, message, and resume bots from the CLI or either dashboard. Bots run as background processes.

shell

$botctl start my-bot--detach$botctl start my-bot--message"check the error logs"$botctl logs my-bot-f$botctl stop my-bot

docs/

* _▾_ Getting Started
* [Overview](https://botctl.dev/docs#overview)
* [Installation](https://botctl.dev/docs#install)

* _▾_ Configuration
* [BOT.md](https://botctl.dev/docs#botmd)
* [Frontmatter](https://botctl.dev/docs#frontmatter)
* [Skills](https://botctl.dev/docs#skills)
* [Workspaces](https://botctl.dev/docs#workspaces)
* [Environment](https://botctl.dev/docs#env)

* _▾_ Interfaces
* [CLI Reference](https://botctl.dev/docs#cli)
* [TUI Dashboard](https://botctl.dev/docs#tui)
* [Web UI](https://botctl.dev/docs#web)

* _▾_ Internals
* [Harness Loop](https://botctl.dev/docs#harness)
*
Read full article → ← Back to Reads