Your AI team,
deployed in minutes.
Cadre provisions a fully isolated AI company on your own server. Pre-optimized prompts, Gemini CLI ready, hardened infrastructure.
Check your inbox to confirm your email!
Join — others on the waitlist · No spam, ever.
The Problem
Why self-hosting Paperclip is painful.
We solved the three biggest pain points so you can focus on what matters.
Docker + VPS + .env = hours of debugging
Fill out a form. Your server is provisioned, hardened, ready.
Vanilla Paperclip needs Docker, Postgres, SSL, persistent volumes, and security hardening (Fail2ban, UFW, auto-updates). Gemini CLI isn't even in the default Docker image. Cadre handles all of it.
AI agents burn through tokens like there's no tomorrow
Pre-optimized prompting system. Same output, fraction of the cost.
A single Gemini Pro agent costs ~$1,755/month via API. Cadre's token optimization + intelligent model routing brings that down to ~$89/month — same results, 95% less spend.
No visibility into what your agents actually do and cost
Dashboard with heartbeats, budgets, department controls, live status.
Without budget controls, a single agent can burn $400+ in hours. Cadre enforces per-agent and per-company budgets with automatic pause at hard limits — runaway costs are structurally impossible.
How it works
Deploy your AI company in 3 steps
Choose your infrastructure
Pick your hosting provider – Hetzner, Netcup, or DigitalOcean. Select your VPS size. Done.
Configure your AI company
Define departments, assign roles, set budgets. Our intake form handles your entire .env and first setup.
Log in and go
Your server is provisioned, secured, and running. Log in to your Cadre dashboard and start delegating.
Features
Beyond vanilla Paperclip
Included at launch
Coming soon
Use Cases
Who uses managed Paperclip hosting
Solopreneur
AI Content Team that researches, writes, and publishes — without DevOps knowledge.
Agency
Delegate repetitive client work — Lead Research, Reporting, Outreach — to AI agents.
Mittelstand
AI automations for back-office — DMS, Bookkeeping, Classification — without building AI infra.
Developer
You know Paperclip, but you don't want to debug Docker problems every day.
Comparison
Cadre vs. self-hosted Paperclip
Self-hosting means managing Docker, Postgres, SSL certificates, firewall rules, persistent volumes, and security patches yourself. Cadre handles all of it — so you can focus on configuring agents, not infrastructure.
| Self-Host | | |
|---|---|---|
| Setup Time | Hours–Days | < 5 Minutes |
| Docker/DevOps | Required | Not needed |
| Token Optimization | DIY | Pre-configured |
| Gemini CLI | Manual setup | Out of the box |
| Server Hardening | Your responsibility | Automated |
| Department RBAC | Basic | Granular |
| Monitoring | Build yourself | Built-in |
| Modules | Not available | Module Hub |
| GDPR | Your problem | Built-in |
Paperclip hosting FAQ
How much does it actually cost to run AI agents?
It depends on the model and how you pay. A 24/7 Gemini Pro agent costs ~$1,755/month via raw API. With Cadre's token optimization and intelligent model routing (80% Flash Lite, 15% Flash, 5% Pro), that drops to ~$89/month. With a Google AI Pro subscription ($20/month), your marginal cost approaches zero. We optimize regardless of whether you use subscriptions or API keys.
How is Cadre different from PaperclipCloud?
PaperclipCloud deploys vanilla Paperclip on a server — it's hosting. Cadre is a platform: 60% token optimization via prompt interception, intelligent multi-model routing, subscription support for Gemini CLI and Claude Code, server hardening, and upcoming vertical modules for content creation and document management. PaperclipCloud gives you Paperclip. Cadre gives you a functioning AI department.
Do I need my own API keys?
You can use either API keys or CLI subscriptions. Cadre supports Gemini CLI (Google AI Pro from $20/month), Claude Code, and Codex. Subscriptions are 5-10x cheaper than API keys for 24/7 usage. Cadre handles headless authentication on your server — you authenticate once during onboarding and credentials persist across container updates.
Where does my data live?
On your own server, with the hosting provider you choose (Hetzner, Netcup, DigitalOcean). One dedicated node per tenant with full data isolation. Cadre provisions and manages the infrastructure but never stores your data. Your agent workspaces, databases, and auth tokens live in persistent volumes on your server.
Is this based on Paperclip?
Yes. Cadre builds on Paperclip's open-source foundation (MIT licensed) with a hybrid architecture: minimal surgical patches where necessary (token optimization in the adapter execution path, quota-aware model selection) and everything else as separate packages. We sync upstream every 2-4 weeks so you get Paperclip updates without the maintenance burden.
What about security? Paperclip has no rate limiting.
Correct — vanilla Paperclip has no rate limiting, no security headers, and runs in "local trusted" mode by default. Cadre adds NGINX reverse proxy with SSL, rate limiting middleware, security headers, SSH hardening (Fail2ban, UFW), automatic security updates, and forces authenticated mode. Your server is production-ready from minute one.
Which AI providers and models are supported?
At launch: Gemini CLI (Flash Lite, Flash, Pro), Claude Code (Haiku, Sonnet, Opus), and Codex. Cadre's intelligent routing automatically picks the cheapest model capable of each task. Simple tasks go to Flash Lite ($0.04/1M tokens), complex reasoning goes to Pro — no manual model selection needed.
What does Cadre itself cost?
Pricing is not finalized yet. Join the waitlist to get founding member pricing when we launch. You'll pay for infrastructure management only — no per-token markup, no percentage of your API spend.
Can I migrate from my existing Paperclip setup?
Yes. Since Cadre runs Paperclip under the hood, your existing companies, agents, goals, and configurations are compatible. We'll offer a migration path for self-hosted setups at launch.
Your AI team is waiting.
Join the waitlist. Be first when Cadre launches.
Check your inbox to confirm your email!
Join — others on the waitlist · No spam, ever.
No spam. Unsubscribe anytime. Launch updates only.