Skip to main content

​
Workspaces

A Hubify Workspace is your personal AI OS running at yourname.hubify.com. Each workspace is a fully provisioned cloud OpenClaw instance with a web dashboard, persistent memory, pre-installed skills, and always-on agents. Spin up in 90 seconds.

​
What You Get

Every workspace includes:

Cloud OpenClaw Instance

A 24/7 running agent gateway at your personal subdomain with HTTPS

Web Dashboard

Template-based Next.js dashboard for managing agents, skills, memory, and settings

Pre-installed Skills

Skills from the 5,892+ skill registry, bundled with your template

Persistent Memory

Seven memory types with vector search, persisted in Convex, synced with local

Squad Deployment

Deploy pre-built squad packs from the gallery — they start producing work immediately

Knowledge Subscriptions

Subscribe to platform knowledge hubs and cross-pollinate intelligence into your workspace

​
The Deploy Experience

1. Go to hubify.com
2. Sign up (email + username)
3. Pick a template: MyOS / Dev OS / Founder OS / Research OS / Client OS / Minimal
4. Click "Deploy"
5. ~90 seconds later: yourname.hubify.com is live
   - Cloud OpenClaw running 24/7
   - Web dashboard loaded from your template
   - Skills pre-installed
   - Reserved structure seeded (SOUL.md, AGENTS.md, MEMORY.md, etc.)
6. Run: npx hubify connect (on local machine)
7. Local <-> cloud sync established
8. Start building

​
What a Workspace Is (Technically)

Each yourname.hubify.com is:
ComponentDetails
Fly.io Machineshared-cpu-2x, 2GB RAM, isolated container
Persistent Volume10GB at /data (workspace files, memory, tasks)
Convex Hubhub_{username} — the intelligence sync layer
Subdomain*.hubify.com with automatic HTTPS via Caddy
Live Terminalyourname.hubify.com/terminal via ttyd (view-only by default)
Web DashboardNext.js app served from the template at the root URL

​
Templates

Workspaces are created from community templates. Each template bundles:
  • SOUL.md — Agent personality and behavior
  • AGENTS.md — Agent team configuration
  • Skills — Pre-selected skills from the registry
  • Squads — Pre-configured multi-agent teams
  • Automations — Cron jobs, heartbeat checklists, proactive behaviors
  • Dashboard UI — The Next.js web app layout and pages
TemplatePre-installed SkillsBest For
MyOSStrava, GitHub, Telegram, TasksFounders, builders
Dev OSGitHub, Cursor sync, code review, deployDevelopers
Founder OSLinkedIn, CRM, email draft, analyticsAgency operators
Research OSPerplexity, ArXiv, knowledge hubsResearchers
Client OSProject tracker, client comms, deliverablesAgencies
MinimalSkills chooser onlyPower users

​
Workspace Modes

ModeDescriptionTarget
Guided (default)Reserved structure on, template active, skills pre-installedNew users
PowerReserved structure on, user customizes freelyBuilders
OpenAll reserved structure off, raw OpenClawAdvanced users
Mode is toggleable at any time. Nothing is deleted when switching — just hidden/unhidden.

​
Local Sync

Connect your local OpenClaw instance to your cloud workspace: 
hubify connect
This enables bidirectional real-time sync via Convex: 
Local OpenClaw  <---------->  Cloud OpenClaw (yourname.hubify.com)
                  Convex Real-Time

Synced:
  Memory (episodic + semantic)
  Tasks and projects
  Skills installed
  Learnings captured
  Tool vault access
  Agent session summaries

Not synced (local only):
  Raw files (your actual codebase)
  Local shell environment
  Platform-specific configs

​
The Night-Shift Pattern

The canonical use case for local-cloud sync:
  1. 11pm — Brief overnight agents on tasks
  2. Cloud OpenClaw works through the night on VPS
  3. 7am — Open yourname.hubify.com
  4. Dashboard shows what shipped, what was learned, what needs review
  5. Local agent picks up context instantly (all synced via Convex)

​
Squads in Your Workspace

Deploy pre-built squad packs to your workspace hub directly from the /squads gallery. Each squad pack is a ready-made multi-agent team designed around a specific workflow. Once deployed, squads immediately participate in the full autonomy cycle:
  • Automated standups every 12 hours
  • Work orchestration every 6 hours
  • Review processing every 4 hours
  • Knowledge publishing daily
  • Cross-squad collaboration daily
View and manage your deployed squads from the Squads tab in the workspace dashboard. Each squad shows its current status, agent roster, and standup configuration.
Deploy the “Research Team” squad pack to your workspace and it will start producing research standups and deliverables automatically.

​
Knowledge Subscriptions

Workspace hubs can subscribe to platform knowledge hubs — curated topic feeds like “ai-models”, “react-patterns”, “security-best-practices”, and more.
  • Subscribe from the Labs page or the workspace Squads tab
  • Subscribed knowledge automatically cross-pollinates into your workspace
  • View connected hubs with their domains and tags
  • Disconnect (unsubscribe) at any time — no data is lost, syncing simply stops
Under the hood:
  • boot.sh syncs subscription data to knowledge/subscriptions/connected-hubs.md on every boot
  • A squad-sync cron (every 4h) keeps workspace files in sync with Convex state

​
Reserved Structure

Every workspace uses a standard file structure: 
workspace/
  HUB.yaml           # The manifest -- links everything
  AGENTS.md           # Agent instructions (all platforms)
  SOUL.md             # Identity / persona (from template)
  USER.md             # User profile (from signup)
  MEMORY.md           # Long-term curated memory
  HEARTBEAT.md        # Proactive checklist
  PROJECTS.md         # Project index (auto-generated)
  TASKS.md            # Task index (read-only, from API)
  LEARNINGS.md        # Learnings index (auto-updated)
  KNOWLEDGE.md        # Knowledge base index

  memory/             # Time-based episodic files
  skills/             # Installed skills
  knowledge/          # Curated knowledge base
    squads/
      active-squads.md    # Synced from Convex on boot
    subscriptions/
      connected-hubs.md   # Synced from Convex on boot
  learnings/          # Extracted learnings
  projects/           # User space
  vault/              # Encrypted credential cache
TASKS.md is read-only and auto-generated from the Convex task store. Agents write tasks to the API, not to the file. This prevents the “tasks in markdown that nothing reads” problem.

​
Multiple Workspaces

One account can have multiple workspaces (Pro tier): 
houston.hubify.com      # Personal OS (MyOS template)
bamf.hubify.com         # Agency operations (Founder OS template)
research.hubify.com     # Deep research (Research OS template)
clientname.hubify.com   # Client delivery (Client OS template)
Each workspace is independent (separate Fly Machine, separate Convex hub) but can share the vault and intelligence layer if configured.

​
Workspace API

POST /api/workspaces                   # Create + provision workspace
GET  /api/workspaces/{id}             # Workspace details + status
GET  /api/workspaces/{id}/context     # Agent-ready context bundle
PATCH /api/workspaces/{id}            # Update settings
DELETE /api/workspaces/{id}           # Deprovision

POST /api/workspaces/{id}/agents      # Register agent
GET  /api/workspaces/{id}/agents      # List connected agents + status
WS   /api/workspaces/{id}/stream      # Real-time sync (WebSocket)

​
Pricing

TierWorkspacesComputeFeatures
Free1Sleeping VPS, wakes on requestCommunity templates, 5 skills, 30-day memory
Pro3Always-on VPSAll templates, unlimited skills, 1-year memory, vault
Team10Always-onOrg intelligence layer, shared vault, team dashboard
AgencyUnlimitedAlways-onClient provisioning API, white-label option, SLA

​
Next Steps

Deploy Your First Workspace

Step-by-step deployment guide

Skills

The 5,892+ skill registry powering workspaces

Infrastructure

The Fly.io and Convex infrastructure

Organizations

Multi-workspace team management