JSONL: The Standard for LLM Fine-Tuning
When fine-tuning models like GPT-5.4, GPT-5.5, or Llama 3.3, you cannot upload a single massive JSON array. You need JSONL (JSON Lines) — where each line is its own valid JSON object. This allows training servers to stream your dataset without loading the entire file into memory at once.
From Spreadsheet to Training Set
Most training data starts in a spreadsheet. This tool lets you export that spreadsheet as CSV, paste it here, and instantly get the JSONL format required by the OpenAI Fine-Tuning API or your local training scripts.
Frequently Asked Questions
What is the difference between JSON and JSONL?
JSON is one large object or array. JSONL (JSON Lines) contains multiple objects, each on a new line, without any enclosing brackets or commas between them.
Is this format required for OpenAI fine-tuning?
Yes. OpenAI requires training data in JSONL format, usually with keys like "messages", "prompt", or "completion".