Cloud Workstation Integration Guide

Overview

CODITECT Cloud Workstations provide fully-managed, cloud-based development environments powered by Google Cloud Workstations. Each workstation includes:

• Code OSS - VS Code-compatible editor
• Claude Dev Extension - AI-powered development assistant
• CODITECT Framework - Pre-installed agents, skills, and commands
• Persistent Storage - 200 GB SSD per workstation
• Full Terminal Access - Bash/Zsh with sudo privileges

---

Architecture

┌─────────────────────────────────────────────────────────────────────────────┐
│                        CODITECT Cloud Architecture                          │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│   User Browser                    CODITECT Backend                          │
│   ───────────                     ────────────────                          │
│   │                               │                                         │
│   │ app.coditect.ai               │ api.coditect.ai                        │
│   │ (React Dashboard)             │ (Django REST API)                      │
│   │       │                       │       │                                 │
│   │       │ ─────── API ──────────┤       │                                │
│   │       │                       │       │                                 │
│   └───────┼───────────────────────┴───────┼─────────────────────────────────┤
│           │                               │                                  │
│           │ WebSocket                     │ Workstation Broker               │
│           │                               │                                  │
│   ┌───────▼───────────────────────────────▼───────────────────────────────┐ │
│   │                     Google Cloud Workstations                          │ │
│   │  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐                    │ │
│   │  │ Workstation │  │ Workstation │  │ Workstation │                    │ │
│   │  │     #1      │  │     #2      │  │     #N      │                    │ │
│   │  │  Code OSS   │  │  Code OSS   │  │  Code OSS   │                    │ │
│   │  │  + Claude   │  │  + Claude   │  │  + Claude   │                    │ │
│   │  └─────────────┘  └─────────────┘  └─────────────┘                    │ │
│   └───────────────────────────────────────────────────────────────────────┘ │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

---

Getting Started

1. Access Your Workstation

1. Log in to app.coditect.ai
2. Navigate to Workstations in the dashboard
3. Click Launch Workstation or select an existing one
4. Wait for provisioning (first launch: ~2 minutes, subsequent: ~30 seconds)

2. Workstation Environment

Your workstation comes pre-configured with:

Component	Version	Description
Code OSS	Latest	VS Code-compatible editor
Node.js	20.x LTS	JavaScript runtime
Python	3.11+	Python runtime with pip
Git	2.40+	Version control
Docker	24.x	Container runtime
Claude Dev	Latest	AI development extension
CODITECT Framework	Latest	~/.coditect symlink

3. First-Time Setup

# Verify CODITECT installation
[ -L ~/.coditect ] && echo "✓ CODITECT installed" || echo "Run setup"

# Check machine ID
cat ~/.coditect/machine-id.json | grep -E '"(machine_uuid|hostname)"'

# Run orientation
/orient

---

Workstation Configuration

Machine Types

Type	vCPUs	Memory	Disk	Use Case
Standard	4	16 GB	200 GB	General development
Performance	8	32 GB	500 GB	Large codebases, ML
Enterprise	16	64 GB	1 TB	Team workloads

Idle Timeout

Workstations automatically stop after inactivity:

Plan	Idle Timeout	Resume Time
Pilot	2 hours	~30 seconds
Pro	4 hours	~30 seconds
Enterprise	8 hours	~30 seconds

---

Integration APIs

Workstation Broker API

The Workstation Broker API manages workstation lifecycle:

Base URL: https://api.coditect.ai/api/v1/workstations/

List Workstations

GET /api/v1/workstations/
Authorization: Bearer <token>

Response:
{
  "workstations": [
    {
      "id": "ws-abc123",
      "name": "my-workstation",
      "status": "RUNNING",
      "machine_type": "e2-standard-4",
      "created_at": "2026-01-10T10:00:00Z",
      "last_used_at": "2026-01-11T14:30:00Z"
    }
  ]
}

Create Workstation

POST /api/v1/workstations/
Authorization: Bearer <token>
Content-Type: application/json

{
  "name": "my-workstation",
  "machine_type": "e2-standard-4",
  "disk_size_gb": 200
}

Response:
{
  "id": "ws-def456",
  "name": "my-workstation",
  "status": "CREATING",
  "estimated_ready_seconds": 120
}

Start/Stop Workstation

POST /api/v1/workstations/{id}/start/
POST /api/v1/workstations/{id}/stop/
Authorization: Bearer <token>

Delete Workstation

DELETE /api/v1/workstations/{id}/
Authorization: Bearer <token>

---

Claude Integration

Claude Dev Extension

The Claude Dev extension is pre-installed and configured:

1. API Key - Managed by CODITECT (no user configuration needed)
2. Context - Automatically includes CODITECT framework context
3. Commands - All / commands available

Using Claude in Workstation

# In terminal
/orient              # Start session
/which "task"        # Find best agent
/agent <name> "task" # Execute with agent

# In editor
Cmd+Shift+P → "Claude: Ask"

---

Persistent Data

What Persists

Location	Persists	Description
/home/user/	Yes	Home directory
/home/user/workspace/	Yes	Project files
~/.coditect/	Yes	Framework data
~/.config/	Yes	VS Code settings
/tmp/	No	Temporary files

Backup & Sync

# Export context
/export my-session

# Sync to cloud
/cx

# Query past work
/cxq "search term"

---

Networking

Outbound Access

Workstations have full outbound internet access:

• GitHub, GitLab, Bitbucket
• npm, PyPI, Docker Hub
• External APIs

Port Forwarding

For local development servers:

# Start dev server
npm run dev -- --port 3000

# Access via workstation URL + port
# https://<workstation-id>.cloudworkstations.dev:3000

---

Troubleshooting

Workstation Won't Start

1. Check license status at app.coditect.ai
2. Verify no resource quota exceeded
3. Try: Stop → Wait 30s → Start

Slow Performance

1. Check machine type (upgrade if needed)
2. Close unused terminals/editors
3. Clear /tmp if disk full

Extension Issues

# Reset Claude Dev
rm -rf ~/.vscode-server/extensions/claude*
# Restart workstation

---

Best Practices

1. Commit frequently - Workstations are ephemeral
2. Use /export - Save context before long breaks
3. Organize workspaces - One project per workstation
4. Monitor usage - Check dashboard for billing

---

Related Documentation

• PILOT-GETTING-STARTED.md - Quick start for pilot users
• PILOT-USER-GUIDE.md - Complete user guide
• MEMORY-MANAGEMENT-GUIDE.md - Context preservation

---

Version: 1.0.0 | Updated: January 11, 2026 Support: support@coditect.ai