SQL to JSON Converter

No. Your SQL is never sent to our servers - all parsing, value coercion, JSON serialization, copy, and download happen directly in your browser on your device. The only network request the tool ever makes is the one you trigger explicitly with URL Import, which goes from your browser to the URL you enter.

Standard INSERT statements with explicit column lists: INSERT INTO table_name (col1, col2, col3) VALUES (val1, val2, val3); Both single-row syntax and multi-row batch syntax (INSERT INTO ... VALUES (...), (...), (...);) are supported. The tool ignores everything outside INSERT statements - so CREATE TABLE, comments, and other DDL are passed over silently.

Unquoted values are auto-typed: NULL becomes null, TRUE/FALSE become booleans, integers become JSON numbers, decimals become JSON numbers, and anything else stays as a string. Quoted string values (in single or double quotes) are always kept as strings, even if they look like numbers. Embedded quotes inside a quoted string are unescaped from the doubled form ('' -> ').

Yes. A statement like INSERT INTO users (id, name) VALUES (1, 'Alice'), (2, 'Bob'), (3, 'Charlie'); produces three JSON objects in the output, one per tuple. The tool walks every (...) tuple after the VALUES keyword until it sees a semicolon or another INSERT.

The Upload picker accepts .sql and .txt files. Drag-and-drop accepts the same. URL Import fetches the file directly from your browser to the target URL - the target site must allow cross-origin requests. You can also type or paste SQL directly into the textarea.

Three: 2 spaces (default), 4 spaces, and Minified (single-line, no whitespace). Pick from the Indent dropdown in the output toolbar after a successful conversion. The setting is saved between visits.

Your settings (indentation) are saved in your browser and restored on the next visit. Your SQL input and generated JSON output are not saved - storing large files on every keystroke would slow the tool down. Re-paste, re-upload, or re-fetch your SQL after a refresh.

The tool is optimized for big files. The full JSON output is always generated and available via Copy and Download, but the on-screen displays are capped: the input textarea is limited to the first 500 lines and the JSON output preview is also limited to the first 500 lines. The actual parsing and JSON generation runs in the background so the page stays responsive.

Those are ignored. The parser scans for INSERT INTO ... VALUES patterns and skips everything else, so a typical SQL dump with CREATE TABLE definitions, comments, and INSERT statements all in the same file works fine - only the INSERTs are converted to JSON.

Scanning the SQL for INSERT patterns, walking column lists and value tuples, coercing types, and serializing to JSON all happen in the background without blocking the page. The spinner overlay animates the whole time, you can still scroll and click other tabs, and the browser never marks the tab as unresponsive. This is what makes the tool work on multi-megabyte SQL dumps.

Both standard single-row syntax (INSERT INTO t (a, b) VALUES (1, 2);) and multi-row batch syntax (INSERT INTO t (a, b) VALUES (1, 2), (3, 4), (5, 6);) are parsed correctly. A batch INSERT with three tuples produces three JSON objects in the output, one per tuple.

Unquoted values are auto-typed: NULL becomes JSON null, TRUE/FALSE become booleans, integers (matching -?\d+) become JSON numbers, decimals (matching -?\d+\.\d+) become JSON numbers, and anything else stays as a string. Quoted string values are always kept as strings, even if they look like numbers.

Both single-quoted ('...') and double-quoted ("...") strings are recognized. Embedded quotes inside a quoted string are unescaped from the standard SQL doubled form ('' -> ') so the JSON output reflects the actual string value. Tabs, newlines, and other whitespace inside quoted strings are preserved.

Column names are extracted from the SQL identifier list and stripped of common quoting characters (backticks, double quotes, square brackets) so the JSON keys are clean. For example, `id`, [id], and "id" all become the JSON key id.

Choose between 2 spaces (default), 4 spaces, or Minified (no whitespace). The Indent dropdown in the output toolbar applies the new indentation immediately without re-parsing the SQL. The setting is saved across visits.

After conversion, a stats line shows how many objects were parsed (one per INSERT row, including each tuple in a multi-row batch INSERT). Useful for sanity-checking the result before copying or downloading.

The parser scans for INSERT INTO ... VALUES patterns and skips everything else. CREATE TABLE definitions, comments (-- and /* */), DROP statements, and other DDL or DML statements are passed over silently. So a typical SQL dump with schema + data all in the same file works without any preprocessing.

Four ways to load SQL: type or paste straight into the textarea, click Upload to pick a .sql / .txt file, drag a file onto the Input area, or click URL Import to reveal a URL field that fetches the file directly from your browser. URL Import requires the target site to allow cross-origin requests.

When you load a large SQL file, only the first 500 lines are shown in the input textarea. The full file is still loaded and used during conversion - the cap is purely a display safeguard to keep the page responsive when the source is multi-megabyte. A notice next to the Input label tells you when the preview is limited.

The output panel shows the generated JSON in a read-only text area. For huge outputs the displayed text is limited to the first 500 lines and a notice next to the Output label tells you the preview is limited along with the full row and column count. Copy and Download always work on the full output - the cap is purely a display safeguard.

When you load a file or fetch a URL, a spinner overlay appears over the Input area with a status message. When you click Convert to JSON, a spinner overlay appears over the Output area until the conversion finishes. This makes it obvious when the tool is working rather than appearing frozen.

Click Copy JSON to copy the entire generated output to your clipboard. The button label briefly changes to "Copied!" for 1.5 seconds as confirmation, then reverts.

Click Download JSON to save the output as a data.json file. The blob is created locally on your device and the temporary object URL is released after the download is triggered.

Your indentation setting is saved in your browser and restored on the next visit. Your SQL input and generated JSON output are not saved - they can be very large, and persisting them on every keystroke would slow the tool down. Re-paste, re-upload, or re-fetch your SQL after a refresh.

Click Clear to wipe the input textarea, the generated JSON, the object count, the URL Import bar, and any error message in a single action. The indentation setting is reset to the 2-space default.