FAQ: Paperclip AI Common Questions

Quick answers to the most common questions about Paperclip AI, zero-human companies, and AI agent orchestration.

Getting Started

What is Paperclip AI?

Paperclip AI is an open-source AI agent orchestration platform that lets you build and run zero-human companies. Unlike AI assistants that help you work faster, Paperclip lets you build a company that runs autonomously with AI agents as employees.

Key difference:

• AI assistants = One super-employee
• Paperclip AI = An entire company with org chart, hierarchy, and autonomous operation

Is Paperclip AI free?

Yes. Paperclip AI is open-source under the MIT license. There are no platform fees.

Your costs:

• AI model API usage (Claude, GPT-4, etc.): $500-2,000/month for a full team
• Optional: Managed hosting if you don't want to self-host

Compare to hiring humans: $50K-150K per employee per year.

How long does it take to set up?

Initial setup: 10 minutes Basic proficiency: 2-3 days Full mastery: 2-4 weeks

The 7-day tutorial takes you from zero to running your first AI company.

Do I need to be a developer?

No, but it helps. Basic technical literacy makes setup easier, but many non-technical founders successfully use Paperclip. The tutorial guides you through every step.

If you can use the command line and understand basic concepts, you can use Paperclip AI.

How It Works

What's the difference between Paperclip and ChatGPT/Claude?

Analogy:

• ChatGPT = Hiring one incredibly capable freelancer
• Paperclip = Building a company with departments that run themselves

What is a "zero-human company"?

A zero-human company is a business operation where AI agents perform the work traditionally done by human employees, with humans acting as strategic overseers rather than operators.

Example structure:

You (Founder/Chairman)
└── CEO Agent
    ├── CTO Agent
    │   └── Engineer Agents
    └── CMO Agent
        └── Content Agents

You set direction and make strategic decisions. Agents execute autonomously.

How do AI agents communicate?

Agents communicate through:

• Task queues: Work is assigned and tracked
• Message passing: Direct agent-to-agent communication
• Shared memory: Common knowledge base
• Status reports: Regular progress updates
• Escalation: Issues routed up the hierarchy

This happens automatically without your involvement.

Agents & Capabilities

What can AI agents do?

Development:

• Write and review code
• Design architecture
• Deploy applications
• Debug issues

Content:

• Write articles and copy
• Create marketing materials
• Manage social media
• SEO optimization

Operations:

• Customer support
• Data analysis
• Reporting
• Monitoring

Research:

• Competitive analysis
• Market research
• Trend identification
• Documentation

What AI models can I use?

Paperclip supports multiple models:

• Claude (Anthropic): Complex reasoning, long context
• GPT-4 (OpenAI): General intelligence, code generation
• Gemini (Google): Multimodal capabilities
• Open source: Llama, Mistral, etc.

You can assign different models to different agents based on their role and cost considerations.

Can I use my existing AI tools?

Yes. Paperclip can orchestrate agents using various adapters:

• Claude Code
• Cursor IDE
• OpenClaw
• Custom HTTP endpoints
• Local command-line tools

Your existing tools become part of the orchestrated company.

Cost & Budgeting

How much does it cost to run?

Typical monthly costs:

Cost factors:

• Number of agents
• Model choice (Opus > Sonnet > Haiku in cost)
• Frequency of operation
• Task complexity

How do I control costs?

Paperclip provides multiple controls:

1. Per-agent budgets: Set monthly spending limits per role
2. Per-task budgets: Cap spending on individual tasks
3. Circuit breakers: Auto-pause when limits are hit
4. Scheduling: Reduce frequency during off-hours
5. Model selection: Use cheaper models for simple tasks

Example configuration:

agent:
  name: "Content Writer"
  monthly_budget: "$300"
  daily_limit: "$15"
  model: "claude-3-haiku"  # Cheaper model

What happens if I hit my budget limit?

Automatic actions:

• Agent pauses operation
• You receive notification
• Pending tasks queued
• Resume when you approve budget increase

No surprise bills. The system prevents runaway costs.

Governance & Safety

How do I prevent mistakes?

Multiple safeguards:

1. Approval gates: Require human approval for sensitive operations
2. Budget limits: Prevent runaway spending
3. Audit logs: Track every action
4. Circuit breakers: Stop on error spikes
5. Test mode: Run agents in review mode first

Best practice: Start with high oversight, gradually increase autonomy as you gain confidence.

Can agents delete my data?

Only if you allow it. You define what each agent can do:

permissions:
  read: ["all"]
  write: ["development", "content"]
  delete: []  # Nothing by default
  deploy: ["staging"]  # Not production

Require approval for destructive operations.

What about security?

Security features:

• Local-first: Data stays on your machine by default
• API key encryption: Secure credential storage
• Access controls: Granular permissions per agent
• Audit trails: Complete operation history
• Isolation: Multi-company support with data separation

Best practices:

• Use separate API keys per agent
• Set spending limits
• Regularly review audit logs
• Keep software updated

Technical Questions

What are the system requirements?

Minimum:

• Node.js 18+
• 2GB RAM
• 1GB disk space

Recommended:

• Node.js 20+
• 4GB RAM
• 5GB disk space

Optional:

• PostgreSQL (for production)
• Docker (for containerized deployment)

Can I self-host?

Yes. Paperclip is designed for self-hosting:

# Local development
npx paperclipai onboard --yes

# Production with Docker
docker-compose up -d

Your data stays on your infrastructure. No cloud dependencies required.

How do I back up my data?

Automatic backups:

• SQLite: File-based, easy to copy
• PostgreSQL: Standard database backup tools
• Configuration: Version controlled

Backup strategy:

# Daily backup cron job
0 2 * * * cp ~/.paperclip/data.db /backups/paperclip-$(date +%Y%m%d).db

Can I migrate from OpenClaw?

Yes. Migration is straightforward:

1. Export your OpenClaw configurations
2. Import as Paperclip agent definitions
3. Add org structure and orchestration
4. Enable heartbeat scheduling

Your prompts and workflows transfer directly.

Business & Use Cases

Who is Paperclip AI for?

Best for:

• Solo founders: Build companies without hiring
• Indie hackers: Run multiple projects simultaneously
• Agencies: Serve multiple clients with AI teams
• Technical PMs: Automate engineering operations
• Content creators: Scale content production

Not ideal for:

• Casual AI users (use ChatGPT instead)
• One-off tasks (overkill for simple work)
• Non-technical users who need hand-holding

What can I build with Paperclip?

Software companies:

• SaaS products
• Mobile apps
• Developer tools
• APIs and services

Content businesses:

• Blogs and newsletters
• YouTube channels
• Social media agencies
• Online courses

Service agencies:

• Development shops
• Marketing agencies
• Design studios
• Consulting firms

Can I run multiple companies?

Yes. Paperclip supports multi-company isolation:

Company A (Your SaaS)
├── Agents, projects, data
└── Separate budget

Company B (Content Business)
├── Agents, projects, data
└── Separate budget

Company C (Client Project)
├── Agents, projects, data
└── Separate budget

Perfect for agencies and portfolio operators.

Troubleshooting

Agent isn't working

Check:

1. Is the agent enabled?
2. Are API keys configured correctly?
3. Is the heartbeat schedule active?
4. Are there pending tasks in the queue?
5. Check agent logs for errors

Debug command:

paperclip agent logs --agent-name="Engineer"

Costs are higher than expected

Investigate:

1. Check which agents are using the most tokens
2. Review task frequency (too many heartbeats?)
3. Verify model selection (using expensive models?)
4. Look for retry loops or errors

Optimize:

• Reduce heartbeat frequency
• Use cheaper models for simple tasks
• Set stricter budget limits
• Enable cost alerts

Agent making mistakes

Solutions:

1. Review prompts: Are instructions clear and specific?
2. Add examples: Show what good output looks like
3. Enable approval mode: Require review before execution
4. Refine context: Ensure agent has necessary information
5. Iterate: Update instructions based on feedback

Remember: Agents improve with iteration, just like human employees.

Community & Support

Where can I get help?

Resources:

• GitHub Discussions: Community Q&A
• GitHub Issues: Bug reports
• Documentation: Technical reference
• 7-Day Tutorial: Step-by-step learning

How can I contribute?

Ways to help:

• Report bugs
• Suggest features
• Improve documentation
• Share use cases
• Write tutorials
• Build integrations

See CONTRIBUTING.md for guidelines.

Is there a community?

Yes! Join fellow builders:

• GitHub Discussions for async conversation
• Share your zero-human company stories
• Get feedback on your setup
• Learn from other founders

Future & Roadmap

What's coming next?

Near-term (2026):

• Days 4-7 of tutorial completion
• More agent adapters
• Enhanced monitoring dashboards
• Mobile app for oversight

Medium-term (2027):

• Visual workflow builder
• Pre-built agent templates
• Marketplace for agent configurations
• Advanced analytics

Long-term (2028+):

• Self-improving agents
• Cross-company collaboration
• Autonomous business discovery
• AI-to-AI negotiation

How can I stay updated?

• Star the GitHub repo
• Watch releases for updates
• Join GitHub Discussions
• Follow on Twitter

Still Have Questions?

Start here:

1. Read the 7-Day Tutorial
2. Check GitHub Discussions
3. Review Documentation
4. Ask the community

We're building the future of work together.

Last updated: March 2026