How to open large JSONL files
Updated November 4, 2026
To open a large JSONL file without a text editor freezing, use a tool that reads part of the file instead of all of it. The JSONL viewer does this in the browser, and we tested it on a 5 GiB file. On the command line, head, sed, jq, ripgrep and DuckDB all handle large files well.
Why editors struggle
Most editors load the whole file into memory and build structures for every line. JSONL files also tend to have very long lines, which many editors handle badly. Because each JSONL line is independent, a tool only needs to read the lines being shown.
Options compared
| Goal | Tool | Notes |
|---|---|---|
| Browse, search and jump around | JSONL viewer | No install; the file stays on your machine |
| See the first records | head -n 5 file.jsonl |
Instant on any size |
| Read one specific line | sed -n '1000p;1001q' file.jsonl |
Stops reading at line 1001 |
| Count lines | wc -l file.jsonl |
Reads the whole file |
| Find text | rg 'needle' file.jsonl |
ripgrep is very fast |
| Filter or reshape records | jq -c 'select(.score > 5)' file.jsonl |
Streams line by line |
| Run SQL queries | DuckDB read_json('file.jsonl') |
Good for aggregates |
| Break the file up | JSONL splitter or split |
Keeps lines whole |
Format a single record
Long records are hard to read on one line. Pipe one through jq:
sed -n '1000p;1001q' file.jsonl | jq .
In the viewer, click a line to see it formatted.
Compressed files
Files ending in .jsonl.gz need decompressing first. To look inside one without writing a copy to disk:
zcat file.jsonl.gz | head -n 5
See the browser benchmark for measured speed and memory use on a 5 GiB file.