Strict by Accident: Your JSON Parser Isn't Broken — It's Answering the Wrong Question

I was importing a JSON file and it blew up. Not a subtle bug — the parser refused the whole thing. The cause was a single extra comma. After deleting it manually, the import worked.

The comma wasn't the problem. The problem was the manual fix, and the all-or-nothing of it: 99% good data thrown out over one byte. So I went looking to see if it was just me.

It wasn't. The same problem shows up again and again — in JavaScript, Python, Ruby, Go, and more. The shape is always the same: the data is there, and the parser throws all of it away because one byte is off-spec.

What developers keep hitting

Almost-JSON from language models. This is the loudest complaint of the last two years. You ask a model for JSON and get something almost right: wrapped in a `json fence, followed by "Let me know if you need anything else!", a trailing comma, a number quoted as a string. One developer stated the reliability problem directly: "It works 95% of the time. The other 5%, you get markdown-wrapped JSON, missing fields, extra commentary… your downstream code crashes." Prompting harder doesn't fix a probabilistic producer; it moves the failures around.