AI Hosting Paperclip-ready AI hosting

Paperclip hosting, one control plane for every agent you run.

Paperclip is an open-source AI orchestration platform: the control plane for turning scattered agents into a structured, company-like workforce. It coordinates agents through adapters to runtimes like Claude Code, OpenClaw, and Hermes. Set goals, assign roles, set budgets, and review task histories from one dashboard. It runs on our AI Hosting plans today.

View AI plans Plan a Paperclip deployment

Paperclip

Agent orchestration layer

Adapters

Claude Code, OpenClaw, Hermes

Budgets

Task and cost tracking

NVMe

Fast AI storage

AI control plane Live on AI Hosting

Paperclip

control plane

Adapters

agent routing

App UI

dashboard

Budgets

task tracking

Compute

Memory

NVMe

$ ai-hosting route paperclip

$ orchestrator init --company=default

$ adapter attach --claude-code --openclaw --hermes

$ budget set --company=default

Why self-host it

Your agents, your infrastructure, your call.

Everything runs on your own server: agent outputs, adapter credentials, and task history never leave your infrastructure. Paperclip coordinates agents 24/7 while your team focuses on higher-level work, with flat pricing instead of per-seat or per-request charges.

View AI plans Talk through a Paperclip build

Private by design

Agent outputs, adapter credentials, and task history stay on your server. Nothing routes through a shared third-party dashboard.

Always-on coordination

Paperclip runs 24/7 coordinating agents with heartbeats, budgets, and task tracking, so nothing stalls between check-ins.

Flat pricing

No per-seat charges, no per-request billing, no escalating costs as your workflows scale.

Model flexibility

Works with Claude, GPT-4, Gemini, and local models. Run Claude Code, OpenClaw, Hermes, and custom adapters in one control plane.

What it's for

What people build with Paperclip.

A control plane for turning scattered agents into a structured, company-like workforce.

Internal operations

Coordinate multiple agents against real operational work instead of running each one by hand.

Autonomous teams

Route work across Claude Code, OpenClaw, and Hermes agents from a single dashboard.

Repeatable automation

Turn recurring tasks into a defined agent workflow with a budget and an owner instead of a one-off script.

Content and knowledge work at scale

Assign agents to research, drafting, and review roles, and review the task history when it's done.

Optional AI stack

Adapters connect Paperclip to the agents you already run.

OpenClaw is a native adapter target. Pair Paperclip with a wider AI stack (interfaces, workflow tools, and model runtimes), all on one plan.

OpenClaw Agent automation

n8n AI Workflow automation

AnythingLLM Team AI workspace

Dify AI app builder

Flowise Visual AI flows

Langflow Agent pipelines

Ollama Model runtime

LocalAI OpenAI-compatible API

vLLM Inference server

Open WebUI Private AI chat

Qdrant Vector database

Weaviate AI-native search

Milvus Large vector search

ComfyUI Image workflows

Whisper Speech to text

JupyterLab AI notebooks

Getting started

From plan to a running control plane.

Requirements scale with how many adapters and concurrent agents you plan to run.

Choose the right plan

Resource needs depend on how many adapters and concurrent agents you're running. Heavier orchestration needs more RAM and cores.

Deploy from GitHub or Docker

Clone the Paperclip repository for full codebase access, or pull the Docker image for simpler updates and cleaner service isolation.

Open the App UI and configure roles

Running a self-hosted model behind an adapter?

Paperclip's control plane itself runs fine on standard plans. It coordinates agents rather than running models directly. If an adapter points at a self-hosted model that needs GPU acceleration, contact us so we can scope it.

Discuss GPU hosting

AI hosting FAQ

Questions before you build on Paperclip?

These answers explain how the orchestration control plane fits with other AI tools, pricing, GPU needs, and support.

Does Paperclip run the AI models itself? No. It coordinates agents through adapters.

Paperclip is the control plane, not the model. It routes tasks, tracks budgets, and monitors heartbeats across adapters. The actual work happens wherever each adapter points, whether that's Claude Code, OpenClaw, Hermes, or a custom integration.

Can Paperclip coordinate Claude Code, OpenClaw, and Hermes together? Yes, in the same company context.

Yes. Paperclip manages task routing, budget allocation, and heartbeat monitoring across multiple adapters at once, so different agents can work under the same goals and budgets.

How is pricing calculated? Flat monthly hosting, not per-seat or per-request.

No per-seat charges, no per-request billing, and no escalating costs as your workflows scale. Trigger as many workflows, task chains, and adapter calls as your plan's infrastructure allows.

Does this include GPU hosting? Not by default. Only needed if an adapter self-hosts a heavier model.

The plans on this page cover the control plane, adapters, and dashboard. If one of your adapters self-hosts a larger model directly, contact us so we can scope a GPU-backed option for that piece.

Can I run this under my own domain? Yes. That is the preferred production path.

You can connect your own domain or subdomain, add SSL, and run the document interface behind your own access rules. This matters for sensitive or internal documents that should not sit on a shared third-party tool.

Start building

Coordinate your agents on hosting that's ready now.

Choose an AI Hosting plan, deploy Paperclip, and wire up adapters for Claude Code, OpenClaw, and Hermes as your agent company grows.

Compare AI plans Talk to sales