# Sample NDJSON file for testing stream processors, free and CC0 1 NDJSON test file of 408.5 KB. A direct URL, the file's SHA-256 and CC0, so curl and CI can fetch it with no signup and no keys. The page: . NDJSON, newline-delimited JSON, puts one complete JSON object on each line. Stream processors, log shippers and bulk APIs prefer it because a reader can handle one record at a time without loading the whole file, and a truncated file loses only its last line. `ndjson/people-1000.ndjson` holds 1,000 people objects from the shared synthetic dataset, one compact object per line, the same records as the JSON, CSV and Parquet people files. That makes it easy to compare a streaming import against a batch import of the same data. Use it to test bulk-import endpoints, streaming parsers, line-based tools such as `jq` and `split`, and upload forms that accept JSON lines. A common bug to look for is a parser that reads the file as one JSON document and fails on the second line. Related formats: [JSON](https://loremfile.dev/json) and [log files](https://loremfile.dev/log). ## Files | Path | Size | MIME | Properties | |---|---|---|---| | [`ndjson/people-1000.ndjson`](https://loremfile.dev/ndjson/people-1000.ndjson) | 408.5 KB (408,525 bytes) | `application/x-ndjson; charset=utf-8` | utf-8, LF, 1,000 lines | The same list as JSON: .