---
sidebar_position: 4
---

# Getting Ready for Agents

The Agent Framework is planned for **H1 2026**, but you can start preparing your infrastructure and workflows today.

This page is based on the "Call To Action" content from the Mycelium Agents page.

## 1. Establish the Networking Layer

Agents will rely on **Mycelium Network** for secure, peer‑to‑peer connectivity.

- Install and use **[Mycelium Network](/network)** on your key devices.  
- Become comfortable with overlay addresses and testing connectivity.  
- Decide which machines, clusters, and locations will be part of your "agent fabric".

## 2. Set Up Compute & Clusters

Use **Mycelium Cloud** and **Virtual Data Centers** as the compute substrate for future agents.

- Deploy clusters via **[Mycelium Cloud Getting Started](/cloud/getting-started)**.  
- Explore **[Kubernetes Basics](/cloud/kubernetes-basics)** and the **[Deployment Tutorials](/cloud/tutorial)**.  
- Design where different workloads should live (personal VDC, shared VDC, edge nodes, etc.).

If you plan to run larger AI workloads, consider capacity with GPUs or high‑memory nodes.

## 3. Decide How You’ll Host Capacity

If you want agents to run on hardware beyond your personal machines, look into becoming a **hoster**:

- Read **[Hosters Overview](/hosters)** to understand the role.  
- Choose hardware via **[Buy or Build a Node](/hosters/buy-node)** and **[Advanced Node Options](/hosters/advanced-nodes)**.  
- Set up your node using **[Set Up Your Node](/hosters/setup)**.  
- Learn about **[Hosting Economics (SPORE)](/hosters/economics)**.

This prepares a foundation where agents can later run on your own nodes or on community capacity.

## 4. Map Your Data & Privacy Requirements

Agents are meant to operate on **your** data, not someone else’s cloud silo.

Begin planning:

- Where your key datasets live today.  
- Which data must stay in specific jurisdictions or networks.  
- Which data could be processed on edge nodes, VDCs, or both.  
- How quantum‑safe and replicated storage (e.g. QSFS) might fit into your architecture.

## 5. Start with Today’s Components

From the call‑to‑action message:

> *“The Agent Framework launches in H1 2026, but the foundation is ready now. Use today’s components — models, storage, compute, and network — to deploy workloads, connect nodes, and prepare for the next generation of distributed AI.”*

Practical steps you can take now:

- Deploy AI workloads on Mycelium Cloud clusters.  
- Use Mycelium Network to connect dev machines, clusters, and edge boxes.  
- Experiment with multi‑node workflows and data flows across environments.  
- Treat these as early prototypes that agents will later formalize and automate.

## 6. Stay Informed

As the Agent Framework evolves, watch for:

- Updated documentation and architecture deep‑dives.  
- Early SDKs, CLIs, or runtimes for prototyping.  
- Example patterns combining agents, VDCs, and QSFS.

Follow official channels (site, docs, community) for announcements.

---

## Summary

You don’t need to wait for the Agent Framework launch to start building:

- Use **network**, **cloud**, **VDC**, and **hosting** pieces today.  
- Organize data and infrastructure so that agents can slot in later.  
- Think in terms of **sovereign, distributed AI** rather than centralized APIs.

When agents arrive, you’ll already have the networking, compute, and data foundations in place.