Skip to content
ED
2026Open source · Library

JSONGuard

Fast, pythonic JSON schema validation, published on PyPI

FIG.04Library Author
Faster than fastjsonschema
116×
Faster than jsonschema
298×
Lines of core code
~250

JSONGuard is a fast, pythonic JSON Schema validator distributed on PyPI. Schemas are defined directly with Python's built-in types and type hints (inspired by sqlmodel), so you skip the boilerplate of formal JSON Schema documents while still getting strict type enforcement, optional fields, and custom validator functions.

The library is roughly 250 lines of optimised code and uses orjson under the hood. Benchmarks place it at ~116× faster than fastjsonschema on basic validation, with depth protection, payload size limits, and either fail-fast or collect-all-errors modes for diagnostics.

What I built

Pythonic schema definitions
Schemas are plain Python dicts using built-in types, Union, and Optional. No separate JSON Schema document or DSL to learn.
Composable custom validators
Any callable that returns True on valid input works as a validator, enabling regex, range, and domain-specific checks directly inside the schema.
Performance-first implementation
Backed by orjson with a tight ~250-line core; benchmarks show ~116× faster basic validation than fastjsonschema and ~298× faster than jsonschema.
Hardening options
Configurable max_size_in_bytes, max_depth, and loosely_typed flags protect against oversized payloads and deeply nested DoS attempts.
Role
Library Author
Timeline
April 2026
Stack
Python, orjson, Type Hints, Pytest, Benchmark tooling
Up next

Sundial Lands

Land-buying and cash-offer platform, built solo end to end

Got something to build? Let's talk.

A full build, an AI integration, or a second opinion on architecture. Most projects start with a thirty-minute call, and I reply to every serious enquiry within a day.

Status
Open to new work
Local time
--:-- UTC+3
Response
Within 24 hours