Metadata-Version: 2.3
Name: json-repair-tool
Version: 0.1.1
Summary: Repair broken JSON and coerce messy LLM outputs into strict Pydantic models.
Keywords: ai,llm,json,pydantic,parsing,openai,rag
Author: Jamessdevops
Author-email: Jamessdevops <subzerotamil2@gmail.com>
Requires-Dist: pydantic>=2.0.0
Requires-Dist: json-repair>=0.25.0
Requires-Python: >=3.10
Description-Content-Type: text/markdown

# Json Repair Tool 🛠️

**Stop your AI pipelines from crashing on bad JSON.**

`json-repair-tool` is a zero-config wrapper that combines fuzzy JSON repair with "soft coercion" for Pydantic models. It is designed specifically for handling the messy, non-deterministic output of Large Language Models (LLMs).

## The Problem

LLMs are bad at strict structured output. They often return:

- Markdown code blocks (` ```json ... `)
- Python booleans (`True` instead of `true`)
- Chatty prefixes ("Here is your data: { ... }")
- "Fuzzy" types ("$100" instead of `100.0`)

## Quick Start

```bash
pip install json-repair-tool
```
