pactown

Fast Start Demo

This example demonstrates pactown’s fast startup capabilities with dependency caching.

What This Shows

• First run: Dependencies are installed and cached (~5-10s)
• Second run: Cached venv is reused (~50-100ms)
• Same deps = same cache: Projects with identical dependencies share the cache

Files

• demo.py - Python script demonstrating fast_run vs regular run
• api/README.md - Sample FastAPI service
• api2/README.md - Second service with same deps (will use cache)

Usage

# Run the demo
python demo.py

# Or use pactown CLI
pactown up fast-start.pactown.yaml

Expected Output

=== Fast Start Demo ===

Run 1 (fresh):
  Creating sandbox...
  Installing dependencies (fastapi, uvicorn)...
  ✓ Started in 5.2s

Run 2 (cached):
  ⚡ Cache hit! Reusing venv
  ✓ Started in 0.08s

Speedup: 65x faster!

How It Works

from pactown import ServiceRunner

runner = ServiceRunner(enable_fast_start=True)

# First run - creates and caches venv
result1 = await runner.fast_run(
    service_id="api-1",
    content=open("api/README.md").read(),
    port=8001,
)

# Second run - reuses cached venv
result2 = await runner.fast_run(
    service_id="api-2",
    content=open("api2/README.md").read(),  # Same deps!
    port=8002,
)

Cache Location

/tmp/pactown-sandboxes/
├── .cache/
│   └── venvs/
│       └── venv_a1b2c3d4/    # Cached: fastapi + uvicorn
├── api-1/
│   └── .venv -> ../.cache/venvs/venv_a1b2c3d4
└── api-2/
    └── .venv -> ../.cache/venvs/venv_a1b2c3d4  # Same cache!

Related Documentation

• Fast Start Guide
• Security Policy

This site is open source. Improve this page.