fleetkit

<p align="center">
  <img src="https://img.shields.io/badge/agents-5-blue" alt="5 Agents">
  <img src="https://img.shields.io/badge/setup-5_minutes-green" alt="5 Min Setup">
  <img src="https://img.shields.io/badge/code_required-zero-orange" alt="Zero Code">
  <img src="https://img.shields.io/badge/license-MIT-lightgrey" alt="MIT License">
</p>

<h1 align="center">⚡ FleetKit</h1>

<p align="center">
  <strong>Get the strategic insight of a $2M executive team — for $1 per decision.<br>Your AI C-suite debates so you don't make $100K mistakes alone.</strong>
</p>

<p align="center">
  <a href="#quick-start">Quick Start</a> •
  <a href="#what-is-fleetkit">What Is It</a> •
  <a href="#see-it-in-action">Demo</a> •
  <a href="#features">Features</a> •
  <a href="#how-it-compares">Compare</a> •
  <a href="#premium-soul-packs">Premium</a>
</p>

---

<p align="center">
  <strong>🎬 Demo Coming Soon</strong><br>
  <em>Your AI executives debating pricing strategy — 4 perspectives, 1 synthesis, 5 minutes</em><br>
  <sub>Until the demo is ready, check out the <a href="examples/brainstorm-example.md">real brainstorm example →</a></sub>
</p>

---

## Quick Start — 5 Minutes to Your First Executive Debate

```bash
# 1. Clone FleetKit
git clone https://github.com/apocys/fleetkit.git

# 2. Copy to your OpenClaw workspace
cp -r fleetkit/ ~/.openclaw/workspace/fleetkit/

# 3. Tell your OpenClaw agent about FleetKit (via Telegram, Discord, etc.)
# "I've added FleetKit to my workspace at fleetkit/. Read fleetkit/workflows/brainstorm.md 
#  — this is how you run multi-agent brainstorms with my AI executive team."

# 4. Run your first brainstorm
# "Brainstorm using FleetKit: Should we price our product as freemium or premium-only?"
```

**That's it.** No code. No Docker. No dependencies. Just files and conversation.

In 2-5 minutes, your AI executives will debate, disagree, and deliver a synthesis with specific recommendations. Hunter (CRO) will push for revenue, Forge (CTO) will flag technical constraints, Echo (CMO) will focus on brand impact. Atlas (COO) will synthesize their debate into actionable strategy.

👉 **[Complete 5-minute setup guide →](QUICKSTART.md)**

---

## See What Happens

You ask: *"Should we price our SaaS at $29/mo or $99/mo?"*

> **💰 Hunter (CRO):** "$99. Period. Your current users already pay — they'll absorb the increase. $29 attracts tire-kickers who churn in month 2."
>
> **🔧 Forge (CTO):** "Hunter's ignoring infrastructure costs. At $29 we need 3x the users but our backend doesn't scale linearly. I'm with him on $99 — but for engineering reasons, not sales reasons."
>
> **📢 Echo (CMO):** "You're both wrong. $99 kills virality. Nobody screenshots a $99 invoice. $29/mo with a $299/year annual plan. The annual plan is the real product."
>
> **📊 Atlas (COO) — Synthesis:** "Consensus: not $29/mo. Tension: monthly vs annual. Recommendation: Launch at $99/mo with $499/year annual (58% savings). Test for 60 days, measure conversion. Echo runs a $29 experiment on a separate landing page."

**That argument took 4 minutes and cost $0.20.** Your last meeting about pricing took 2 hours and decided nothing.

---

## Before FleetKit vs After FleetKit

**BEFORE:** Tuesday's 3-hour meeting. Hunter wants to move fast. Sarah wants more data. Mike thinks it cheapens the brand. Nothing decided. Meeting rescheduled for Thursday.

**AFTER:** You ask FleetKit the same question. 5 minutes: four perspectives, genuine disagreement, a synthesis with specific next steps. Decision made. Tuesday afternoon freed up.

*This is what happens when your team argues efficiently.*

---

## What Is FleetKit?

FleetKit is an **open-source starter kit** for running multi-agent AI teams on [OpenClaw](https://openclaw.com).

Meet your AI C-suite. They argue about everything. That's the point:

- **5 AI executives** with distinct personalities that genuinely disagree
- **Workflows** for brainstorms, standups, and decision escalation
- **Protocols** that produce real insight, not AI slop
- **A dashboard** to see your fleet's activity

### What FleetKit is NOT

- ❌ Not a framework — it's a template. Fork it, edit it, make it yours.
- ❌ Not autonomous agents — your AI executives advise, YOU decide.
- ❌ Not a coding project — everything is configured through SOUL.md files and YAML.
- ❌ Not magic — it's structured prompting done right. The magic is in the personalities.

---

## The Team

| Agent | Role | Personality | Catchphrase |
|-------|------|-------------|-------------|
| 👑 **Chief** | CEO | Decisive, vision-driven, allergic to theater | *"Does this still matter in 5 years?"* |
| 📊 **Atlas** | COO | Meticulous, process-obsessed, evidence-first | *"Has this been documented?"* |
| 💰 **Hunter** | CRO | Aggressive, numbers-forward, speed-obsessed | *"What's the revenue impact?"* |
| 🔧 **Forge** | CTO | First-principles, optimization-obsessed, dry humor | *"Delete it. If nobody notices, it wasn't needed."* |
| 📢 **Echo** | CMO | Creative, provocative, empathetic, bold | *"Nobody cares about features. What's the feeling?"* |

They're designed to **disagree**. Hunter pushes for speed, Forge pushes back on technical debt. Echo wants bold branding, Atlas wants documented processes. The friction is the feature.

---

## See It In Action

Here's an excerpt from a real brainstorm on revenue strategy:

> **💰 Hunter (CRO):** Digital product packs — fastest path to $10K/mo. Developer market pays premium. Revenue by week 4. $5-15K/mo potential. *Focus everything here.*
>
> **🔧 Forge (CTO):** "Revenue by week 4" is aggressive bordering on delusional. Has Hunter actually mapped the funnel? Product → landing page → traffic → conversion → payment. That's 4 things that all need to work simultaneously.
>
> **📢 Echo (CMO):** *This is our iPhone moment.* Every customer becomes an influencer. 9/10 viral potential. The community becomes a self-reinforcing flywheel.
>
> **🔧 Forge:** "iPhone moment" is inspiring and zero percent actionable. Convert the vision into a spec and I'll build it in a week. Until then, it's poetry.
>
> **📊 Atlas (COO) — Synthesis:** Kill the SaaS idea (unanimous). Digital packs have the highest ceiling. Ship 5 packs in 2 weeks, launch on Product Hunt week 3. Revenue target: 150-350 sales/month = $4K-$25K.

**[See the full example →](examples/brainstorm-example.md)**

---

## Features

- 🧠 **5 Distinct Agent Personalities** — CEO, COO, CRO, CTO, CMO with unique decision frameworks
- 🔄 **3-Round Brainstorm Protocol** — Positions → Debate → Synthesis
- 📋 **Daily Standups** — Automated morning briefs delivered to Telegram/Discord
- ⚖️ **Decision Escalation** — Structured protocol for when agents disagree
- 📊 **Fleet Dashboard** — See agent activity, costs, and status at a glance
- ⏰ **Cron Jobs** — Ready-to-use configs for automated standups and retros
- 📝 **Memory System** — Agents build context over time through MEMORY.md files
- 🎛️ **Fully Configurable** — YAML config for models, temperatures, schedules, delivery
- 💰 **Cost Tracking** — Built-in budget alerts and per-operation cost estimates
- 🔌 **Zero Dependencies** — Just text files. No code, no Docker, no build step.

---

## How It Compares

| Feature | FleetKit | CrewAI | AutoGen | LangGraph |
|---------|----------|--------|---------|-----------|
| Setup time | **5 minutes** | 30+ minutes | 1+ hours | 2+ hours |
| Code required | **None** | Python | Python | Python |
| Agent personality | **Deep SOUL files** | Basic roles | Minimal | Minimal |
| Structured debate | **3-round protocol** | Sequential | Chat | Graph |
| Decision escalation | **Built-in** | Custom code | Custom code | Custom code |
| Daily standups | **Built-in** | Custom code | Custom code | Custom code |
| Dashboard | **Included** | Third-party | Third-party | Third-party |
| Cost per brainstorm | **~$1** | ~$2-5 | ~$3-8 | ~$2-5 |
| Learning curve | **Read a README** | Learn Python SDK | Learn Python SDK | Learn Python SDK |

**FleetKit's philosop

[truncated…]