JSONL Merger

Why cat file1.jsonl file2.jsonl can corrupt your data

Concatenation looks like the obvious way to merge JSONL — and it works right up until one file lacks a trailing newline. Then the last record of one file and the first record of the next fuse into a single invalid line:

# file1.jsonl ends WITHOUT a trailing newline:
{"id": 1}
{"id": 2}⏎missing

# file2.jsonl:
{"id": 3}

# cat file1.jsonl file2.jsonl produces:
{"id": 1}
{"id": 2}{"id": 3}   ← one broken line: parse error, two records lost

This tool checks every input's final byte and inserts the missing newline before joining — the merged output is always valid, regardless of how the inputs were saved. If you must use cat in a pipeline, the defensive form is: awk 1 file1.jsonl file2.jsonl (awk re-terminates every line).

How to Merge JSONL Files

Select or drag and drop multiple JSONL files into the upload area. Each file is read entirely in your browser — no data is sent to any server. Files can be in .jsonl, .ndjson, .json, or .txt format.

Once uploaded, you can reorder files using the up/down arrows to control the merge order. Enable deduplication to remove duplicate lines based on their full content, or enable JSON validation to skip lines that are not valid JSON.

The merged output appears in the preview panel showing the first 100 lines. Download the full merged file as a .jsonl file or copy it to your clipboard. Statistics show the total line count, file size, and any duplicates or invalid lines removed.

Merge semantics, precisely

Order

Files are concatenated in the order you add them, and record order within each file is preserved. Merging is append-only — records are never reordered, rewritten, or reformatted.

Boundary repair

A missing trailing newline at the end of any input is fixed automatically (see the example above). Blank lines between records are dropped, since some consumers reject them.

Deduplication: exact line

Optional. Two records count as duplicates only when their raw lines are byte-identical — {"a":1,"b":2} and {"b":2,"a":1} are NOT deduplicated, because key order makes the lines differ even though the JSON values are equal. Normalize with the formatter's sort-keys option first if you need value-level dedup.

Round-trip with the splitter

Merging is the exact inverse of splitting: split a file into N chunks, merge them back in order, and the result is byte-identical to the original (with dedup off). This property makes split → process in parallel → merge a safe pipeline pattern.

What is JSONL Merging?

JSONL merging is the process of combining multiple JSONL (JSON Lines) files into a single file. Each input file contains one JSON object per line, and the merger concatenates all lines from all files into one output file while preserving the order you specify.

This is essential when working with datasets split across multiple files — such as training data shards, daily log exports, or chunked API responses. Merging consolidates them into a single file for easier processing, analysis, or uploading to platforms that expect a single JSONL input.

Common Errors & Solutions

Frequently Asked Questions