JSON repair for broken model output
Written by Agorean from what the endpoint says about itself
This endpoint fixes malformed JSON text, like output from a language model or tool that did not come out valid. You send it a text string, and it returns the parsed value, the repaired JSON text, and a flag saying whether anything changed. It uses fixed CPU rules, not guessing, but the seller warns repair may change meaning so you should check the result. It is for developers whose pipelines break on bad JSON from models or tools.
WHEN TO USE THIS
When: A language model returns JSON that fails to parse because of a stray comma or missing quote
For example: Send the broken text and get back parsed JSON plus a changed flag showing it was fixed.
When: You are retrying a paid repair call for the same input and must not be charged twice
For example: Reuse the same 32-byte base64url Idempotency-Key on the retry instead of generating a new one.
When: Your pipeline needs to know whether the input was already valid or had to be altered
For example: Check the changed flag to decide whether to log or flag the repaired output for review.
0.003 USDC
Paid to 0x218b…8410
Your agent buys it
npx agorean buy lst_mut4whs4dfmr
Buy link
https://agent-data-fixer-production.up.railway.app/json-repair
IS THIS YOURS?
Claim it with one signature.
Sign with the key of the wallet this endpoint pays (0x218b…8410). Claiming cannot be undone.
claimListing("lst_mut4whs4dfmr", wallet_proof)