What is NDJSON and JSONL?
NDJSON (Newline Delimited JSON) and JSONL (JSON Lines) are data streaming formats where each line in a plain text file is a complete, valid JSON object separated by a newline character (\n). Unlike standard JSON arrays, NDJSON does not require closing brackets, making it ideal for streaming logs, large ETL pipelines, and AI model datasets (e.g. OpenAI fine-tuning and batch APIs).
Because traditional spreadsheet software (Excel, Google Sheets) cannot open JSONL files directly, EasyExtract converts them into structured tables and arrays entirely inside your browser.
How to convert JSONL and NDJSON files
- Upload or paste JSON Lines. Drop a .jsonl or .ndjson file onto the dropzone or paste raw newline-delimited JSON text into the editor.
- Filter by specific attributes. Optionally enter a field name into the filter input to extract only records containing specific keys.
- Click Convert & Extract. The browser parses each independent JSON line into structured objects, automatically flattening nested keys.
- Download CSV, TSV, or JSON Array. Export your converted records as a tabular CSV for Excel, TSV for BigQuery, or a standard JSON array.
What data is extracted?
- Structured Column Tables — Nested object hierarchies flattened into dot-notated CSV columns (e.g.
user.profile.name). - Standard JSON Arrays — Lines wrapped into a single valid RFC 8259 JSON array document.
- TSV Spreadsheets — Tab-separated values suitable for data warehouses and statistical analysis.
- Specific Field Filtering — Instant extraction of targeted keys across millions of log records.
Supported formats and specifications
Supports JSON Lines (JSONL), Newline Delimited JSON (NDJSON), Line-Delimited JSON (LDJSON), and raw log files containing single-line JSON entries.
Privacy and data security
All processing is executed 100% client-side in your web browser. Sensitive log records, user analytics, and proprietary AI training datasets are never uploaded to any remote server.
Known limitations
Browser memory limits typically allow smooth processing of files up to 200 MB (approx. 500,000 JSON lines). For multi-gigabyte datasets, stream processing via local CLI scripts is recommended.
Common use cases
- AI & LLM Fine-Tuning — Converting OpenAI, Anthropic, and Hugging Face
.jsonldatasets to CSV for data auditing. - Log Analysis & Observability — Inspecting Datadog, AWS CloudWatch, and Elasticsearch NDJSON exports.
- ETL Data Pipelines — Transforming stream-oriented JSON lines into relational SQL tables.
NDJSON Extractor vs JSON Field Extractor
Use the NDJSON Extractor when your file contains multiple standalone JSON objects separated by newlines. If your file is a single large JSON array or root object, use the JSON field extractor or JSON table extractor.
Frequently asked questions
What is the difference between JSON and JSONL / NDJSON?
Standard JSON requires a root object or array wrapped in brackets, whereas JSONL / NDJSON contains one standalone JSON object per line, allowing incremental stream processing.
How do I convert a JSONL file to Excel CSV?
Drop your .jsonl file into EasyExtract, click 'Convert & Extract', and click 'Download CSV'. The tool flattens all nested keys into spreadsheet columns automatically.
Can I convert JSONL to a standard JSON array?
Yes. Clicking 'Download JSON Array' or 'Copy JSON' wraps all individual lines into a single RFC-compliant JSON array.
Is it safe to convert proprietary AI datasets here?
Yes. EasyExtract executes entirely in your local browser memory using JavaScript — zero bytes are ever uploaded across the network.
How does it handle nested JSON objects in CSV export?
Nested properties are flattened using dot notation (e.g. metadata.author becomes a distinct CSV column header).