Free AI data tool · Formats
Chat dataset validator
Quick answer
Paste JSONL where each record has a messages array. Pick a profile and see every issue by record and message.
Limit: 10 MiB or 25,000 records
Your sample is processed in this browser. File contents are not uploaded.
Supported input
JSONL, one conversation per line, up to 10 MiB or 25,000 records.
How it works
Generic schema: each record is an object with messages[]; each message has role (string) and content (string, null, or an array of parts that each have a type). A system message, if present, must be first. Profiles decide allowed roles, whether empty content is allowed and whether an assistant reply is required (anywhere or as the last message).
Limitations
- This is a generic schema, not a check against any specific provider's current format.
- Passing structure checks says nothing about training quality, accuracy or rights.
Questions
- What format does a chat fine-tuning dataset use?
- Commonly one JSON object per line with a messages array, where each message has a role (system, user, assistant) and content. Providers differ in details, so check their current docs too.
- Is my file uploaded?
- No. The file is read and processed in this browser tab, and its contents are not sent to SourceX or anyone else. Closing the tab or pressing Reset clears it.
Learn the format
Related tools
JSONL validatorFind the exact lines that break your JSONL file and export only valid records.Dataset quality checkerMeasure completeness, exact duplicates and text length in a CSV, JSON or JSONL sample.Dataset requirements / RFP builderWrite a clear dataset requirements brief and download it as Markdown or JSON.
Request data in this format
Need licensed business data delivered in this format? Describe it and we'll scope it.
Request data in this format