In partnership with
What Is Paperclip?
Paperclip lets you build and manage AI agents. Agents are small programs that research, write, and review work automatically. You connect them into workflows so they work like a team. A web dashboard lets you control everything. It runs on your computer, a cloud server, or inside Docker.
Who uses it: Content teams, developers, business owners, and anyone wanting automated AI workflows with budget and approval controls.
What you need: A terminal, Git, Node.js 18+, npm, a database (SQLite or PostgreSQL), optionally Docker, and an OpenAI API key.
Full Workflow (Start → End)
Docker install
↓
Paperclip install
↓
Dashboard open
↓
Account create
↓
Company create
↓
Goal set
↓
1 agent add
↓
API key add
↓
Simple task test
↓
Error fix
↓
More agent add
↓
Budget set
↓
Full workflow run
↓
Monitor systemQuick Setup: Choose Your Path
Path 1 — Local Computer (Easiest)
# Install Node.js from nodejs.org, then verify
node -v && npm -v
# Clone the project
git clone https://github.com/paperclip/paperclip.git paperclip-app
cd paperclip-app
npm install
# Create .env file with these lines
DATABASE_URL="file:./dev.db"
OPENAI_API_KEY=sk-your-key-here
# Initialize database
npx prisma migrate dev --name init
# Start
npm run devOpen http://localhost:3000. Create your admin account. Done.
Path 2 — VPS Server (Intermediate)
Rent Ubuntu 22.04 VPS (2 vCPU, 4 GB RAM). SSH into it:
ssh root@your-server-ip
apt update && apt upgrade -y
apt install -y curl git ufw postgresql postgresql-contrib
# Install Node.js 20
curl -fsSL https://deb.nodesource.com/setup_20.x | bash -
apt install -y nodejs
# Setup PostgreSQL
sudo -i -u postgres
createuser --interactive # Name: paperclipuser, superuser: n
createdb paperclipdb -O paperclipuser
exit
# Deploy Paperclip
cd /opt
git clone https://github.com/paperclip/paperclip.git paperclip-app
cd paperclip-app
npm install
nano .env # Fill DATABASE_URL, OPENAI_API_KEY, NEXTAUTH_URL, NEXTAUTH_SECRET
npx prisma migrate dev --name init
npm run build
npm run start
# Open firewall
ufw allow 3000/tcp
ufw enableVisit http://your-server-ip:3000.
Path 3 — Docker (Quick & Clean)
mkdir paperclip-docker && cd paperclip-docker
wget https://raw.githubusercontent.com/paperclip/paperclip/main/docker-compose.yml
# Create .env with DATABASE_URL, OPENAI_API_KEY, NEXTAUTH_SECRET, POSTGRES credentials
docker compose up -d
docker compose exec app npx prisma migrate dev --name initOpen http://localhost:3000.
Your reach is rented. And landlords evict.
One algorithm update. One policy change. One bad quarter for a platform that isn't yours. The audience you spent years building disappears overnight.
beehiiv is what happens when you stop renting and start owning. A list that's yours. Revenue that compounds. Growth tools built in from day one.
30% off your first 3 months with code LIST30. Start building today.
First Steps in Dashboard
Create admin account on the registration page.
Create a workspace (like “My AI Team”).
Verify API key under Settings → API Status. Should say “Connected”.
Building Agents: Beginner → Expert
Beginner: Single Agent
Name: Simple Researcher
Model: gpt-3.5-turbo
Instructions: “You are a research assistant. Search the web and return a short summary with three key points.”
Tool: Web Search ON
Test in the chat box.
Intermediate: Multi-Agent Workflow
Create two agents:
Researcher — gathers bullet points.
Writer — turns bullets into a blog post (model: gpt-4o).
Build a workflow:
Start → Researcher → Writer → End.
Run manually with a topic like “Benefits of solar energy”.
Expert: Budget, Approval, Monitoring
Budget: Set a monthly hard limit (e.g., $50) in Workspace Settings → Budget. Agents stop when reached.
Approval: Add an Approval node between Researcher and Writer. Workflow pauses until a human clicks “Approve”.
Monitoring: Check the Logs page. Set email alerts when cost reaches 80% of budget under Settings → Notifications.
Complete AI Content Team Workflow
Research Agent (gpt-4o + web search)
Approval Step (human checks research)
Writer Agent (gpt-4o, friendly tone)
Reviewer Agent (gpt-4o, grammar and fact-check)
Final Approval
Set per-run budget ($3). Enable notifications. You now have a production-ready content pipeline.
Common Problems and Quick Fixes
Problem | Fix |
|---|---|
| Reinstall Node.js, restart terminal |
Port 3000 in use | Set |
Database connection error | Check DATABASE_URL; ensure PostgreSQL is running |
API key invalid | Verify key starts with |
White dashboard screen | Clear cache; check |
Agent gives poor output | Improve instructions; use gpt-4o; increase max_tokens |
Unexpected high cost | Set strict budget; use gpt-3.5-turbo for drafts |
Docker container restarts |
|
Beginner go-to fix: Read the last line of the error message. Restart the service. Double-check .env for typos.
Security Essentials
Never commit
.envto Git. Add it to.gitignore.Use strong passwords for admin and database.
On VPS, use a reverse proxy (Nginx) with SSL, or Cloudflare Tunnel for secure access.
If a key leaks, revoke it immediately and create a new one.
Roadmap Summary
Beginner: Local install → one agent → manual runs → learn prompts.
Intermediate: VPS/Docker → two-agent workflow → budget limits → email notifications.
Expert: Approval steps → advanced models with tools → monitoring → automated backups → HTTPS → scaling.
Best Practices
Start with cheap models (gpt-3.5-turbo), test, then upgrade.
Always set a budget, even for testing.
Keep old prompt versions when editing agents.
Monitor usage weekly.
Back up database and
.envregularly.
You now have the complete path from zero to a fully automated AI team. Test, build, and scale at your own pace. Happy automating!
Arnold Schwarzenegger has a newsletter.
Yeah. That Arnold Schwarzenegger.
So do Codie Sanchez, Scott Galloway, Colin & Samir, Shaan Puri, and Jay Shetty. And none of them are doing it for fun. They're doing it because a list you own compounds in ways that social media never will.
beehiiv is where they built it. You can start yours for 30% off your first 3 months with code PLATFORM30. Start building today.