Open Source  •  19 AI Employees  •  MIT License

startup-agents

Build a $100K startup with AI employees.

Product. Design. Engineering. Marketing.
An entire company in your terminal.

19 AI Employees
4 Departments
1 Command

One coding agent
isn’t enough.

Before

Most AI tools give you

one coding agent

It writes code. That’s it. No product thinking, no design perspective, no marketing strategy. No shared context across tasks.

  • No product thinking
  • No design perspective
  • No marketing strategy
  • No shared context
After

startup-agents gives you

an entire company

Product, design, engineering, and marketing — all working together with shared memory so every decision stays in context.

  • Product Manager defines your MVP
  • Designer builds the brand
  • Engineer ships the code
  • Marketer plans the launch

19 AI employees.
4 departments.

Each agent has specialized skills, workflows, and deliverables — just like a real hire.

Product

5 agents
  • Product Manager Strategy, roadmap, MVP definition
  • Product Analyst User research, data insights
  • Sprint Planner Backlog, velocity, prioritization
  • ASO Specialist App store optimization
  • Monetization Strategist Pricing, revenue modeling

Design

4 agents
  • Brand Guardian Brand identity, voice, strategy
  • UI Designer Visual design, component libraries
  • UX Architect CSS architecture, layout systems
  • Logo Designer Logomarks, wordmarks, icon systems

Engineering

5 agents
  • Frontend Developer React/Next.js, performance, a11y
  • Backend Developer APIs, databases, architecture
  • Mobile Developer Expo SDK 54, iOS & Android
  • AI Engineer LLM integration, prompt engineering
  • DevOps Engineer CI/CD, infrastructure, deployment

Marketing

5 agents
  • Content Strategist Blog, copy, storytelling
  • SEO Specialist Keywords, on-page SEO, backlinks
  • Social Media Manager Platform strategy, community
  • Growth Hacker Viral loops, acquisition channels
  • Email Marketing Specialist Sequences, lifecycle, deliverability

Three steps to
your AI company.

Clone the repo

Get the full startup-agents framework with all 19 role definitions, agent wrappers, and the shared memory system.

$ git clone https://github.com/easonwang00/Startup-Agents.git
$ cd Startup-Agents

Run claude

Claude Code reads CLAUDE.md automatically on startup. All 19 agents become available instantly via @mention or natural language.

$ claude
> CLAUDE.md loaded — 19 agents ready

Just talk to your team

Describe what you need. The right agent activates, executes with deep expertise, and writes its decisions to shared memory so every other agent stays in sync.

You “Let’s build a tool for freelancers to track time and send invoices.”
Product Manager Validating market, defining MVP, writing to .memory/product.md…

The brain that keeps everyone in sync.

Every agent reads from and writes to .memory/ — so your Engineer knows what the Designer decided, and your Marketer knows what the PM prioritized.

company.md Mission, business model, target market
product.md Features, tech stack, architecture
brand.md Colors, fonts, tone of voice, logo
decisions.md Decision log with reasoning & dates
progress.md What’s done, in progress, up next
metrics.md KPIs, revenue targets, growth numbers

Up and running
in 60 seconds.

$ git clone https://github.com/easonwang00/Startup-Agents.git
$ cd Startup-Agents
$ claude
CLAUDE.md loaded — 19 agents ready
Shared memory initialized
Ready. Just talk to your team.

Example prompts

“I have a startup idea — let’s build it”
Product Manager activates
“Design the brand for our product”
Brand Guardian activates
“Build the landing page”
Frontend Developer activates
“Set up pricing and monetization”
Monetization Strategist activates
“Plan our Product Hunt launch”
Growth Hacker activates
View on GitHub

Stop building alone.
Hire your AI team today.

Free. Open source. MIT licensed. No subscriptions, no lock-in.

Get startup-agents on GitHub