ndcorder

ndcorder / outputguard

Public

Validate, repair, and retry LLM structured outputs. 13 repair strategies for common JSON malformations, JSON Schema validation, and retry-with-feedback prompts.

github.comndcorderoutputguard ai anthropic cli developer-tools json
15
0
100% credibility
Found May 12, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

OutputGuard is a Python library that automatically validates, repairs malformed structured data like JSON or YAML from AI language models, and generates retry prompts for invalid outputs.

How It Works

1
📰 Discover the fixer

You're frustrated because AI chats keep giving you messy lists or data that doesn't fit right, then you find this handy tool that automatically cleans it up.

2
📥 Add it easily

You bring the tool into your project with a quick download, like adding a new app to your phone.

3
📋 Set your rules

You describe what perfect data looks like for you, such as names as words and ages as numbers.

4
✨ Clean it magically

You give it the broken AI output, and it checks against your rules, fixes common mistakes like extra marks or wrong quotes, and hands back neat data.

5
🔄 Smart retries

If it's still not quite right, it creates a helpful follow-up question to ask the AI again until it gets it perfect.

6
📊 Handle lots at once

For big batches of data, you process many pieces together and get a summary of what got fixed.

🎉 Perfect data every time

Now your AI outputs are always clean and reliable, saving you hours of manual fixes and letting you focus on building cool things.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 15 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is outputguard?

outputguard is a Python library that validates, repairs, and retries structured outputs from LLMs, tackling common JSON malformations like trailing commas, markdown fences, and truncated objects. It applies 13 repair strategies, checks against JSON schemas, and supports YAML, TOML, or Python literals via simple APIs or CLI commands like `outputguard validate file.json -s schema.json --repair`. Developers get clean data back fast, without custom parsers.

Why is it gaining traction?

It stands out by fixing real-world LLM messes—tested on 288 models with 100% success—while offering guarded generation wrappers, batch validate & repair for evals, and retry prompts that feed errors back to models. CLI handles quick fixes on outputs, with reports showing strategies applied and confidence scores, beating ad-hoc regex scripts.

Who should use this?

AI engineers integrating LLMs for JSON APIs or tools needing reliable structured extraction. Evals teams batch-validating LLM outputs against schemas. Backend devs wrapping any LLM client to auto-repair common malformations without provider lock-in.

Verdict

Worth adding to LLM pipelines for its thorough tests (2,001 passing) and docs, despite 15 stars and 1.0% credibility signaling early maturity—use in non-critical paths first. Solid for validate & repair workflows today.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.