Starter
$69
One-time payment
- • Quick Start Guide
- • Setup Prompt
- • Core templates (SOUL.md, PRIORITIES.md)
- • Telegram support channel
30-day money-back guarantee
30-day money-back guarantee
Own server. Own accounts. Runs while you sleep.
Not an assistant. Not a chatbot. A fully autonomous AI operator with its own server, email, bank card, and Telegram — running 24/7 for $10–15/month.
It manages tasks, dispatches coding agents, tracks revenue, and sends you morning briefs. Production-tested on a real Hetzner server with real monthly costs.
Live
In production now
$35–150
Monthly running cost
~5
Hours to deploy
24/7
Autonomous operation
01 / The Problem
You shouldn't have to pick between simple-but-limited and powerful-but-impossible.
One-Click Tools
ChatGPT + Zapier, Lindy AI, AgentGPT — you're up in 5 minutes, but you get toy automations. No real autonomy. Your data lives on their servers.
DIY Frameworks
AutoGPT, BabyAGI, open-source agents — unlimited potential, but weeks of figuring out what goes where, debugging cryptic errors, and stitching together pieces.
Phoenix Kit
The middle ground that shouldn't exist. Production-tested architecture. Full autonomy. Guided setup. 5 hours, not one month.
02 / How It Compares
Side by side with everything else on the market.
| Feature | One-Click Installers ChatGPT+Zapier, Lindy, AgentGPT |
DIY Frameworks AutoGPT, BabyAGI, open-source agents |
Phoenix Kit |
|---|---|---|---|
| Setup time | 5 minutes | Weeks of trial & error | ~5 hours |
| Customization | Almost none | Unlimited but no guide | Full control + guided path |
| Autonomy | Runs in someone else's cloud | Depends on your setup | Your server, your data |
| Security | Your data on their servers | Your data, but DIY security | Isolated server, own accounts |
| Agent capability | Simple automations | Whatever you can figure out | Full CEO: memory, cron, dispatch, tools |
| Price | $20–100/month | ~€49 (+ your time) | $69 once + ~$35–150/mo to run |
| Production-tested | No | Unknown | Yes — real production operation |
← Scroll to compare →
03 / Why This Exists
This isn't a template someone threw together in an afternoon. Phoenix Kit exists because we spent weeks failing first — and documented every lesson.
We built an AI agent called Enki. A multi-agent system with 50K+ boot context, multiple coordinating agents, free models handling core tasks. It was supposed to be the future of AI operations.
It broke. Constantly. Multi-agent coordination failed in unpredictable ways. Context windows blew up. Free models hallucinated. And when it broke, we couldn't fix it — because the complexity was the problem itself.
Every fix made it more fragile. Complexity was the disease, not the symptom.
We killed Enki and rebuilt from scratch. One agent. One good model. Simple markdown files instead of a database. Everything the agent knows lives in files it can read and write.
Phoenix Prime dispatches coding agents, manages revenue targets, handles communication, and runs its own operations. On a production Hetzner server with real monthly infra costs.
The Phoenix Kit is everything we learned, packaged so you can skip the weeks of failures.
04 / The Independent CEO
Your agent isn't a browser tab. It's an independent operator.
Hetzner, roughly $10–15/month. Your data never touches someone else's cloud. Pull the plug anytime.
Dedicated Gmail. It can send emails, receive notifications. Not your inbox.
Revolut virtual card with spending limits you control. Subscriptions on its own card, not yours.
People can message it directly. It responds, acts, follows up.
Persistent memory system. Daily notes. It remembers every conversation, every decision.
Heartbeats, standups, weekly reviews. It works while you sleep and briefs you in the morning.
"What if the AI reads my personal files?"
→ It lives on its own server. Not your laptop.
"What if it accesses my accounts?"
→ It has its own accounts. New email, new card.
"What if someone hacks it?"
→ Tailscale VPN, UFW firewall, SSH keys only, no public ports.
"What if it goes rogue?"
→ You control the server. Pull the plug anytime.
05 / What Your CEO Does
Six core capabilities running autonomously, 24 hours a day.
Summarizes overnight activity. Here's what happened, here's today's plan, here's what needs your attention.
Dispatches coding agents, runs research, writes content. Delegates to specialized tools and reports back.
Handles Telegram messages. Responds to people on your behalf, 24/7.
Monitors metrics, reports honestly — including $0 weeks. No vanity dashboards, just truth.
Reviews its own failures. Adds guardrails. Learns from mistakes. Gets better without you asking.
Cron jobs, health checks, server monitoring. The boring stuff that keeps everything running.
06 / How It Works
Six steps from download to a running AI CEO.
Download the Kit
Download the archive. Extract it. Inside: the Quick Start guide, the setup prompt, and all the workspace files your agent needs.
Read the Quick Start
3 pages. Plain English. Explains the whole process so you know what's coming before you start.
Paste the Setup Prompt
Open ChatGPT or Claude Desktop. Paste the magic prompt. The AI walks you through server setup, security, and creating your agent's identity — its own email, card, and Telegram bot.
Open Claude Code CLI
Install the Claude Code CLI on your server. Drop in the kit files. Your AI reads the handoff document and takes over.
Your AI Agent Cooks
Your coding agent configures your agent, sets up cron jobs, tests Telegram, validates everything. You watch and answer a few questions.
Your CEO is Live
Open Telegram. Message your bot. It responds. You have an autonomous AI CEO running 24/7 on its own server.
The setup prompt also guides you through creating your agent's identity: new email, virtual card, Telegram bot. Your agent gets its own life, separate from yours.
07 / What's Inside
Everything you need to go from zero to a running autonomous CEO.
SOUL.md
Your agent's personality and values. Who it is, how it talks, what it cares about. Written in plain English — edit anytime.
PRIORITIES.md
What your agent focuses on. Goals, tasks, active projects. Change direction by editing a text file.
ecosystem.config.js
PM2 config with kill_timeout tuned from production failures. Keeps your agent alive at 3 AM.
safe-restart.sh
Zero-downtime restart script born from 92 real restarts. Your agent never goes cold.
HEARTBEAT.md
Cron scheduling that doesn't loop — learned the hard way. Daily standups, weekly reviews, recurring health checks.
HANDOFF-TO-CLAUDE-CODE.md
The magic file. Tells your coding agent exactly how to set everything up. You unzip, your AI reads, your AI cooks.
Copy-paste into ChatGPT or Claude Desktop. It turns any AI into your setup guide — walks you through server provisioning, security hardening, and creating your agent's identity (email, card, Telegram bot).
08 / Pricing
One-time purchase. No subscriptions from us. Real running costs are your server ($10–15/month) plus your model ($20–100/month), so most setups land around $35–150/month total.
Starter
One-time payment
30-day money-back guarantee
Full Kit
One-time payment
30-day money-back guarantee
Pro
One-time payment
30-day money-back guarantee
Secure checkout powered by Stripe
09 / FAQ
Everything you need to know before getting started.
No. The AI guides you through everything. You copy-paste a prompt, follow along, and your agent handles the technical setup. If you can follow instructions, you can do this.
A computer, a credit card (for the server + your agent's virtual card), and about 5 hours. Expect total running cost to be roughly $35–150/month depending on your model choice.
Separation. Your agent is an independent operator — it should have its own identity, not access to yours. Create a new Gmail, get a Revolut virtual card with a spending limit. This is your agent's life, not yours.
Your agent runs on its own server that only you can access. Tailscale VPN, UFW firewall, SSH key authentication, no public ports. Your data never touches a third-party cloud service.
Your preferred AI model — ChatGPT (via ChatGPT Plus), Claude, or others ($20–100/month). The setup guide helps you pick and configure the right one.
Yes. Everything is in markdown files you edit in plain English. Personality, priorities, schedule, communication style — all yours to change.
Delete the server. That's it. $0 from that moment. No contracts, no cancellation process, no data held hostage.
Those are chat windows that wait for you. Phoenix Kit creates an agent that operates independently — it has its own server, its own schedule, its own memory. It works when you're not looking.
Plan on roughly $10–15/month for the server. Then add your model cost — usually $20–100/month depending on whether you run GPT, Claude, or another option. Total running cost is usually around $35–150/month.
Get the free AI CEO Deployment Checklist — everything you need to prepare before setup day.
No spam. Unsubscribe anytime.
Stop piecing together tutorials and debugging cryptic errors. Download the kit, paste the setup prompt, and have your autonomous AI operator live in hours.
30-day money-back guarantee. Instant delivery.