⚖️ Probatio

Put your data to the proof.

About

Probatio is a modern, maintained data validation library for Python. The model is simple: a schema is data, and data describes data. You compose schemas from plain Python types, dicts, lists, and a handful of small helpers, then call the schema with a value to validate it.

It is a drop-in for voluptuous: the same public API, so you can swap the import and keep your existing schemas. Probatio is a clean-room reimplementation (the same API written fresh, not a fork), which means it can be actively maintained, ships under a clean MIT license, and is free to fix bugs and improve its internals without inheriting old decisions. See ADR-001 for the full reasoning, and Credits and inspiration for the lineage.

It does not stop at parity. Probatio clears voluptuous's own backlog (cross-field rules, dataclass and TypedDict schemas, network and format validators, errors that carry a path and suggest the key you meant), and it is held to the bar you would want from a library that loads untrusted config. See Why trust it.

It is also fast for pure Python: roughly two to three times voluptuous's speed on a representative config schema, and about seven times once a schema compiles itself. The numbers, and how to reproduce them, are in the performance reference.

Probatio is pure Python: no native extension, nothing to build at install time. Install it and import it. Requires Python 3.12 or newer.

Installation

pip install probatio

Or with uv:

uv add probatio

Usage

Define a schema, then call it with a value to validate it. A valid value comes back (possibly normalized); an invalid one raises Invalid.

from probatio import Schema, Required, Optional

schema = Schema(
    {
        Required("name"): str,
        Optional("port", default=8080): int,
    }
)

schema({"name": "app"})
# {'name': 'app', 'port': 8080}

When a value does not match, the error carries a path to the offending value, however deep it sits:

from probatio import Schema, Invalid

schema = Schema({"server": {"port": int}})

try:
    schema({"server": {"port": "nope"}})
except Invalid as err:
    print(err)
    # expected int at 'server.port'

Why trust it

It is a config-loading library, so the real question is whether you would feed it untrusted input. The evidence, not the adjectives:

• voluptuous 0.16.0 test suite: runs against Probatio, with every divergence a deliberate, documented deviation. voluptuous's own authors' notion of the contract, checked.
• Home Assistant config_validation: the full suite passes, with voluptuous swapped out for Probatio.
• 100% line and branch coverage, type-checked under both mypy and ty, in CI.
• Fuzzed on every untrusted-input surface. The first fuzzing pass found hundreds of exception leaks on hostile input. All fixed; none since.
• Safer than the original: a built-in validator only ever raises Invalid, never a raw exception, and that is enforced, not hoped for.

Migrating from voluptuous

Probatio aims to be a drop-in replacement. In most cases the migration is a single import change:

# Before
from voluptuous import Schema, Required, Optional, All, Any, Coerce, Invalid

# After
from probatio import Schema, Required, Optional, All, Any, Coerce, Invalid

The markers (Required, Optional, Remove, Extra), combinators (All, Any), helpers (Coerce, Range, In, Length, Match, and friends), and errors (Invalid, MultipleInvalid) all behave the way they do in voluptuous. See the migration guide for the current compatibility status.

Documentation

Full documentation lives at probatio.frenck.dev: getting started, the migration guide, and the API reference.

Changelog & Releases

This repository keeps a changelog using GitHub's releases functionality. The format of the log is based on Keep a Changelog. There is intentionally no CHANGELOG.md file: the GitHub Releases are the changelog.

Releases are based on Semantic Versioning, and use the format of MAJOR.MINOR.PATCH. In a nutshell, the version will be incremented based on the following:

• MAJOR: Incompatible or major changes.
• MINOR: Backward-compatible new features and enhancements.
• PATCH: Backward-compatible bugfixes and package updates.

Contributing

This is an active open-source project. We are always open to people who want to use the code or contribute to it.

Using AI tools to help is fine, but you must review and understand everything you submit. Please read our AI Policy first; autonomous agents are not allowed, and unreviewed AI output will be closed.

Before you start, read the contributing guide, the code of conduct, and the security policy. Bugs and feature requests go to the issue tracker.

Thank you for being involved.

Credits and inspiration

Probatio owes its design to voluptuous by Alec Thomas. The API and the "schema is data" validation model are its inspiration. Probatio reimplements them fresh, no code copied, so the result can be maintained and MIT licensed, but the original idea is theirs and the credit belongs to them.

voluptuous itself drew on earlier work, and it is only fair to pass that on: Validino as its major influence, with lighter nods to jsonvalidator and json_schema. Probatio stands on that same lineage.

Authors & contributors

The original setup of this repository is by Franck Nijhof.

For a full list of all authors and contributors, check the contributors page.

License

MIT License

Copyright (c) 2026 Franck Nijhof

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.