Frequently Asked Questions

General

What is Pyvorin?

Pyvorin is a Python acceleration platform. The thin client sends compatible Python code to a remote compiler that produces native machine code, while unsupported code falls back to CPython transparently.

Do I need to change my Python code?

Often no. If your code uses supported constructs, you can run it directly. If it contains unsupported features, Pyvorin falls back to CPython so it still works. For best performance, refactor hot paths to use primitive types and simple loops.

What is the difference between pyvorin and pyvorin-thin?

pyvorin-thin is the lightweight client that offloads compilation to the cloud. The full pyvorin package includes a local native compiler and is typically used in self-hosted or enterprise deployments.

Licensing & Billing

How is the licence validated?

The thin client sends a hash of your licence key to the Pyvorin API. The raw key is never transmitted in plain text.

What happens if my licence expires?

The thin client continues to function but falls back to CPython for all executions. You will not experience crashes; you simply lose native compilation until the licence is renewed.

Is there a free tier?

Please check the pricing page on the main site for current plans and trial options.

Performance

How much faster will my code run?

Speedups are workload-specific. ETL pipelines often see 10–1,000× speedups on warm runs, while JSON parsing may see 5–40×. Microbenchmarks can show higher numbers, but real-world speedups depend on input size, type stability, and compilation tier.

Does compilation overhead cancel out the speedup?

For short scripts run once, yes. For long-running jobs, repeated function calls, or large data sets, the warm-run speedup far exceeds the one-time compile cost.

Can I run benchmarks myself?

Yes. Use pyvorin-thin benchmark script.py. It measures CPython and Pyvorin timings, checks correctness, and writes an auditable event.

Security & Privacy

Does Pyvorin upload my source code?

Not by default. Source upload is disabled. The thin client sends a fingerprint (hash) of your code. You can enable raw source upload with --include-source after setting PYVORIN_THIN_SEND_SOURCE=1.

Where is my data stored?

Usage events and benchmark results are stored in your local queue directory. They are transmitted to Pyvorin servers only when you opt in and connectivity is available. Telemetry can be disabled entirely.

Compatibility

Does Pyvorin support Python 3.12?

The thin client requires Python 3.9+. Compatibility with the latest Python versions is tracked in the release notes.

Does Pyvorin work on macOS and Windows?

The thin client runs on any platform that supports Python. Compilation targets are platform-specific; the default target is linux_x86_64. Contact support for other targets.

Can I use Pyvorin with Docker?

Yes. Install pyvorin-thin inside your container image and configure the API URL and licence key at runtime via environment variables.