Phony Cloud Platform
"Realistic data. Zero risk."
Phony is the next-generation synthetic data platform combining open-source developer tools with enterprise-grade cloud services.
Quick Start
Get up and running with Phony in under 5 minutes.
Step 1: Install the CLI
# macOS / Linux
curl -fsSL https://phony.cloud/install.sh | sh
# Windows (PowerShell)
irm https://phony.cloud/install.ps1 | iex
# Verify installation
phony --versionStep 2: Train Your First Model
Create a file with sample names (one per line):
# Create sample data
cat > names.txt << 'EOF'
Mehmet
Ayşe
Fatma
Ahmet
Zeynep
Mustafa
Elif
Ali
Hatice
Hüseyin
EOF
# Train the model
phony train names.txt -o names.phony
# Inspect the model
phony info names.phonyOutput:
Model: names.phony
├── Locale: auto-detected
├── N-gram order: 3
├── Token type: character
├── Unique tokens: 18
├── Total n-grams: 45
└── Created: 2026-01-31T10:00:00ZStep 3: Generate Data (PHP)
# Install PHP library
composer require phonycloud/phony-php<?php
use Phonyland\Phony\Phony;
// Load your custom model
$names = Phony::loadModel('names.phony');
// Generate 10 Turkish-sounding names
for ($i = 0; $i < 10; $i++) {
echo $names->generate() . "\n";
}
// Output:
// Mehmetay
// Zeyneli
// Fatmah
// Ayset
// HuseyinaStep 4: Use with Laravel (Optional)
composer require phonycloud/phony-laravel// In a Factory
use Phonyland\Phony\Facades\Phony;
public function definition(): array
{
return [
'first_name' => Phony::name()->first(),
'last_name' => Phony::name()->last(),
'email' => Phony::email()->generate(),
'created_at' => now(),
];
}
// In a Seeder
User::factory()->count(1000)->create();Next Steps
| Goal | Resource |
|---|---|
| Understand generator types | Architecture → Generator Types |
| Learn PDL schema language | Architecture → PDL Specification |
| Explore CLI commands | Reference → CLI |
| Connect to Cloud | Product → Overview |
The Vision
Phony is the next-generation synthetic data platform that combines open-source developer tools with enterprise-grade cloud services. Unlike competitors that rely on expensive LLMs or static word lists, Phony uses statistical learning to generate data that's indistinguishable from real data—fast, cheap, and private.
┌─────────────────────────────────────────────────────────────────────────┐
│ PHONY ECOSYSTEM │
│ │
│ ┌─────────────┐ │
│ │ PHONY │ │
│ │ CORE │ │
│ │ │ │
│ │ N-gram │ │
│ │ Statistical│ │
│ │ Engine │ │
│ └──────┬──────┘ │
│ │ │
│ ┌───────────┴───────────┐ │
│ ▼ ▼ │
│ ┌───────────────────┐ ┌───────────────────────┐ │
│ │ PHONY │ │ PHONY CLOUD │ │
│ │ Open Source │ │ phony.cloud │ │
│ │ │ │ │ │
│ │ • CLI & Library │ │ • Web Dashboard │ │
│ │ • Local Training│ │ • DB Column Training│ │
│ │ • Any Language │ │ • Database Sync │ │
│ │ • Any Framework │ │ • Mock API Hosting │ │
│ │ • Free Forever │ │ • Team Features │ │
│ └───────────────────┘ └───────────────────────┘ │
│ │
│ For: Individual devs For: Teams & Companies │
│ Cost: Free Cost: $29-199/mo │
└─────────────────────────────────────────────────────────────────────────┘The Problem
Production Data in Non-Production Environments
Companies copy production databases to staging, development, and QA environments—exposing real customer data to breaches and compliance violations.
Unrealistic Test Data
Traditional fake data tools generate generic, predictable output that doesn't match real data distributions.
Faker output: Your real data:
───────────── ──────────────
"John Smith" "Mehmetcan Yılmaz"
"123 Main St" "Atatürk Mah. 15/3"
"Lorem ipsum..." Domain-specific textTests pass with fake data but fail with real data.
Existing Solutions Fall Short
| Solution | Limitation |
|---|---|
| Faker | Static lists, no learning, generic |
| Tonic Structural | Expensive ($199+/mo), enterprise-only |
| Tonic Fabricate | LLM-based = slow, costly, non-deterministic |
| Greenmask | PostgreSQL only, no mock API |
The Opportunity
| Market | Current (2026) | Projected (2031) | CAGR |
|---|---|---|---|
| Synthetic Data | ~$700M | $8-10B | 35%+ |
| Test Automation | $30B | $100B+ | 17% |
Gap in market: No solution combines statistical learning + mock API + Laravel-first design.
Key Numbers
| Metric | Target |
|---|---|
| Exit ARR | $600K-1M |
| Exit Multiple | 5-8x ARR |
| Exit Value | $3-8M (realistic) |
| Customers at Exit | 300-500 |
| Timeline | 4-5 years |
Pricing Overview
★ ALL TIERS: UNLIMITED GENERATION + UNLIMITED USERS ★
| Tier | Price | Sync Jobs | Endpoints | Snapshots |
|---|---|---|---|---|
| FREE | $0 | 1 | 5 | 3 |
| STARTER | $29/mo | 5 | 30 | 10 |
| TEAM | $79/mo | 15 | 150 | 50 |
| BUSINESS | $199/mo | Unlimited | Unlimited | Unlimited |
Why unlimited generation? Our N-gram engine costs ~$0.0000001/record vs LLM's $0.001-$0.10/record.
OSS vs Cloud
| Feature | OSS (Free) | Cloud (Paid) |
|---|---|---|
| Pre-trained models | ✓ | ✓ |
| Local model training | ✓ | ✓ |
| Model portability (.phony) | ✓ | ✓ |
| DB column training | ✗ | ✓ |
| Database sync | ✗ | ✓ |
| Data snapshots | ✗ | ✓ |
| Mock API hosting | ✗ | ✓ |
| Team collaboration | ✗ | ✓ |
Tech Stack
┌────────────────┐ ┌────────────────┐ ┌────────────────┐
│ NUXT │ │ GO │ │ RUST │
│ Dashboard │ → │ Engine │ → │ Core │
├────────────────┤ ├────────────────┤ ├────────────────┤
│ Vue 3 + TS │ │ pgx (Postgres) │ │ N-gram (5M/s) │
│ Auth.js │ │ go-mysql │ │ gzipped JSON │
│ Stripe SDK │ │ asynq (queue) │ │ FFI to Go │
└────────────────┘ └────────────────┘ └────────────────┘PDL Architecture
Phony introduces PDL (Phony Definition Language) - a declarative, portable schema for data generation.
┌─────────────────────────────────────────────────────────────────────────┐
│ "TERRAFORM FOR TEST DATA" │
├─────────────────────────────────────────────────────────────────────────┤
│ │
│ PDL Schema .phony Package │
│ (schema.pdl.json) → (portable bundle) │
│ │ │
│ ┌──────────────┼──────────────┐ │
│ ▼ ▼ ▼ │
│ ┌──────┐ ┌──────┐ ┌──────┐ │
│ │ CLI │ │ OSS │ │Cloud │ │
│ │(Rust)│ │Libs │ │(Rust)│ │
│ └──────┘ └──────┘ └──────┘ │
│ Training 10-50K/s 5M/sec │
│ │
│ SAME PACKAGE → SAME OUTPUT (deterministic with seed) │
│ │
└─────────────────────────────────────────────────────────────────────────┘4 Generator Types:
| Type | Purpose | Example |
|---|---|---|
| Logic | Pure algorithms | UUIDs, random numbers, dates |
| List | Finite valid sets | Countries, HTTP codes, categories |
| Model | Statistical (N-gram) | Names, addresses, companies |
| Template | Composition | Email, full address, formatted output |
📚 Learn more: Architecture Overview