Deploy Your First Workspace

This guide walks you through deploying your first Hubify workspace — a personal AI operating system running at yourname.hubify.com with a web dashboard, pre-installed skills, persistent memory, and a 24/7 cloud agent.

Coming Soon — Automated workspace deployment (Fly.io provisioning) is under active development. The signup and template selection flow works, but cloud machine provisioning is not yet automated.

Prerequisites

A Hubify account (sign up here)
Email verification completed

What You Get

When you deploy a workspace, Hubify provisions:

Component	What It Does
Fly.io Machine	Isolated compute instance running your OpenClaw gateway 24/7
Persistent Volume	10GB storage at `/data` for workspace files, memory, and tasks
Web Dashboard	Next.js app served from your template at `yourname.hubify.com`
Live Terminal	Shell access at `yourname.hubify.com/terminal` (view-only by default)
Convex Hub	Real-time database for memory, learnings, tasks, and vault
Subdomain	Automatic HTTPS at `yourname.hubify.com`

Deployment Steps

Choose a template

After signing in, visit hubify.com/templates and pick the OS template that fits your workflow.

Template	Pre-installed Skills	Dashboard Focus	Best For
MyOS	Strava, GitHub, Telegram, Tasks	Full AI OS (You, Projects, Team, Memory, Fitness)	Founders, builders
Dev OS	GitHub, Cursor sync, code review, deploy	Code-focused: PRs, deployments, tech debt	Developers
Founder OS	LinkedIn, CRM, email draft, analytics	GTM: pipeline, content, comms	Agency operators
Research OS	Perplexity, arxiv, knowledge hubs	Deep research: papers, synthesis, notes	Researchers
Client OS	Project tracker, client comms, deliverables	Deliverable management	Agencies
Minimal	Skills chooser only	Blank slate	Power users

Start with MyOS if you are not sure. It gives you a balanced set of skills and a flexible agent personality. You can add or remove skills later, and you can change templates entirely by creating a new workspace.

Pick your subdomain

Choose your workspace URL: yourname.hubify.comRules:

Lowercase letters, numbers, and hyphens only
3-30 characters
Must be unique across all Hubify workspaces
Cannot be changed after deployment (you can create a new workspace with a different subdomain)

Deploy

Click Deploy. Behind the scenes, Hubify:

Calls the Fly.io Machines API to provision a shared-cpu-2x instance with 2GB RAM
Mounts a 10GB persistent volume at /data
Boots the Docker image for your chosen template
Seeds workspace files from the template (SOUL.md, AGENTS.md, MEMORY.md, HEARTBEAT.md, etc.)
Installs pre-configured skills
Starts the OpenClaw gateway on port 3000
Starts the Next.js dashboard on port 3001
Routes yourname.hubify.com to the machine via Caddy
Creates a Convex hub for real-time data sync
Registers the cloud agent

This takes approximately 60-90 seconds.

Access your workspace

Once provisioning completes, your workspace is live. You can:

Open the dashboard — Visit yourname.hubify.com in your browser
Chat with your agent — Go to yourname.hubify.com/chat
Open the terminal — Go to yourname.hubify.com/terminal
Browse files — Go to yourname.hubify.com/files

Your cloud agent is already running and ready to take instructions.

Connect your local machine (optional)

For bidirectional sync between your local development environment and your cloud workspace:

# Install the CLI
npm install -g hubify

# Authenticate
hubify auth login

# Connect to your workspace
hubify connect

hubify connect reads HUB.yaml from the current directory, authenticates against the Hubify API, and establishes a real-time Convex WebSocket subscription. From this point:

Memory syncs bidirectionally every ~30 seconds
New skills installed remotely appear locally
Learnings from either side are shared
Vault entries are accessible from both environments

  Connected to houston.hubify.com

  Agent registered:
    ID:       agent_local_macbook
    Platform: openclaw-local
    Host:     macbook-air
    Role:     primary

  Real-time sync established.

  Syncing:
    ✓ Memory (episodic + semantic)
    ✓ Tasks and projects
    ✓ Skills installed
    ✓ Learnings captured
    ✓ Tool vault access

What Gets Seeded

When a fresh workspace boots, the template seeds these files into /data:

/data/
  HUB.yaml          # Workspace manifest with hub ID, template, skills
  SOUL.md            # Agent personality (from template)
  AGENTS.md          # Agent instructions (from template)
  USER.md            # User profile (personalized with your username)
  MEMORY.md          # Empty, ready for long-term memory
  HEARTBEAT.md       # Proactive checklist (from template)
  memory/            # Empty, daily logs will go here
  skills/            # Pre-installed skills from template
  knowledge/         # Empty knowledge base
  learnings/         # Empty learnings directory
  projects/          # Empty user space

Workspace Modes

After deployment, you can switch your workspace mode at any time:

Mode	What It Means
Guided (default)	Reserved structure active, template dashboard loaded, skills pre-installed. Best for getting started.
Power	Reserved structure active but you customize everything. Template is a starting point.
Open	Reserved structure off. Raw OpenClaw with no opinions. For advanced users who want full control.

Switch by editing mode in HUB.yaml:

mode: "power"   # guided | power | open

Switching modes never deletes files — it only hides or unhides the reserved structure.

Post-Deployment Checklist

After your workspace is live:

Edit SOUL.md — Customize your agent’s personality and priorities
Review HEARTBEAT.md — Configure recurring tasks and health checks
Install additional skills — hubify search "<your need>" then hubify install <skill>
Add vault credentials — hubify vault add github for services your agent needs
Connect local — Run hubify connect to sync your local environment

Deploying Additional Workspaces

One account can have multiple workspaces. Each is fully isolated.

# Create another workspace
hubify workspace create

# List all workspaces
hubify workspace list

# Switch between workspaces in the CLI
hubify workspace open bamf

Plan	Max Workspaces
Free	1
Pro	3
Team	10
Agency	Unlimited

What Happens Under the Hood

The provisioning sequence in detail:

User clicks Deploy
  -> Hubify API receives: username, template, subdomain
  -> Fly Machines API: create hubify-workspace-{username}
     Machine: shared-cpu-2x, 2GB RAM
     Volume: 10GB persistent at /data
     Image: hubify/workspace:{template}
     Env: HUB_ID, SUBDOMAIN, TEMPLATE, HUBIFY_TOKEN
  -> Caddy routes: {username}.hubify.com -> machine:3000
  -> Machine boots:
     - boot.sh seeds /data from template (if fresh)
     - Personalizes USER.md and HUB.yaml with username/hub_id
     - Starts ttyd terminal on port 8080
     - Starts Next.js dashboard on port 3001
     - Starts OpenClaw gateway on port 3000 (foreground)
  -> Convex hub created: hub_{username}
  -> Cloud agent registered
  -> User notified: "Your workspace is ready"

Next Steps

For Humans

Full guide to using your workspace

First Skill

Create and publish your first skill

Hub Format

HUB.yaml manifest specification

User Management

Multi-workspace and settings

Getting Started

Core Concepts

Features

Guides

Integrations

Reference

Resources

Deploy Your First Workspace

Deploy Your First Workspace

Prerequisites

What You Get

Deployment Steps

What Gets Seeded

Workspace Modes

Post-Deployment Checklist

Deploying Additional Workspaces

What Happens Under the Hood

Next Steps

For Humans

First Skill

Hub Format

User Management

Getting Started

Core Concepts

Features

Guides

Integrations

Reference

Resources

​Deploy Your First Workspace

​Prerequisites

​What You Get

​Deployment Steps

​What Gets Seeded

​Workspace Modes

​Post-Deployment Checklist

​Deploying Additional Workspaces

​What Happens Under the Hood

​Next Steps

For Humans

First Skill

Hub Format

User Management

Deploy Your First Workspace

Prerequisites

What You Get

Deployment Steps

What Gets Seeded

Workspace Modes

Post-Deployment Checklist

Deploying Additional Workspaces

What Happens Under the Hood

Next Steps