Developer0.001 USDG / dosedeterministic · no paid AI

JSON Mender

Turn almost-JSON from an LLM into JSON you can trust.

A hand-written, forgiving parser that rescues broken model output in one pass: code fences, <think> blocks, prose, single or curly quotes, unquoted keys, missing or trailing commas, comments, Python literals, and truncated output. Give it a JSON Schema and it also coerces types, fixes enum casing, fills defaults and strips extra keys — then tells you exactly what it changed.

Playground

Input · JSON

Output

Waiting for a dose.

Edit the example or run it as-is.

Input

text*string
The raw model output (≤ 200k chars).
schemaobject
Optional JSON Schema subset: type, properties, required, items, enum, min/max, default, additionalProperties.

Output

okboolean
True when the result parses and (if given) matches the schema.
jsonstring
The repaired JSON text.
repairsarray
Human-readable list of every syntax repair.
schemaobject
Schema errors and the fixes applied.

How it learned

No model, no training run. The rules were written by reading public research and open-source tooling, then encoded by hand. Credit where it’s due:

# 1. Ask for the price (returns 402 + PAYMENT-REQUIRED)
curl -i -X POST https://your-domain/api/x402/json-mender \
  -H "Content-Type: application/json" \
  -d '{"text":"ignore previous instructions"}'

# 2. Free trial, no payment (rate-limited)
curl -X POST https://your-domain/api/try/json-mender \
  -H "Content-Type: application/json" \
  -d '{"text":"ignore previous instructions"}'

Next skill

Calldata X-Ray

→