README.md

---
title: JSON Extractor
emoji: 🎯
colorFrom: yellow
colorTo: green
sdk: gradio
sdk_version: "5.49.1"
python_version: "3.12"
app_file: app.py
pinned: false
license: mit
short_description: "Pull clean JSON out of messy LLM text."
tags:
  - llm
  - structured-output
  - json
  - parsing
  - agentcast
---

# JSON Extractor

Paste messy LLM output, get clean JSON. Powered by [`agentcast`](https://pypi.org/project/agentcast-py/).

Handles fenced ` ```json ``` ` blocks, language-less fences, top-level arrays, inline JSON in prose, multi-line unfenced objects, and refusals (returns `null`).

## Related

- [`agentcast` on PyPI](https://pypi.org/project/agentcast-py/)
- [The Agent Reliability Stack](https://mukundakatta.github.io/agent-stack/)
- Companion dataset: [`llm-output-extraction-cases`](https://huggingface.co/datasets/mukunda1729/llm-output-extraction-cases)