JSON ⇄ CSV-konverterare

Convert JSON to CSV and CSV to JSON in your browser

This JSON to CSV converter turns a JSON array of objects into a clean, spreadsheet-ready CSV — and converts CSV back into structured JSON — without uploading a thing. Paste your data, drop in a file, or load the sample, then flip the direction toggle to go either way. The result appears live on the right with a preview table, so you can check the columns before you Copy or Download.

It handles the fiddly parts of CSV for you: values that contain commas, quotes, or line breaks are quoted and escaped on the way out, and parsed back correctly on the way in, so your data round-trips between JSON and CSV without corruption.

Flatten nested JSON to CSV

Real-world JSON is rarely flat, and that is where most converters fall over. Switch on Flatten nested and this tool expands nested objects and arrays into dot-notation columns: an address object becomes address.city and address.country, while a roles array becomes roles.0 and roles.1. That lets you flatten nested JSON to CSV in one click instead of writing a script.

Tools like ConvertCSV.com and csvjson.com handle simple flat JSON well but require manual pre-processing for nested structures. Here, you switch one toggle. Prefer to keep a sub-object together? Leave flattening off and each nested value is stored as JSON text inside a single cell, which converts straight back to the original structure when you reverse the direction.

Open your CSV in Excel or Google Sheets

Spreadsheets cannot read JSON directly, but every major spreadsheet app reads CSV without any plugin or conversion step. The fastest path from JSON to Excel is to convert here and open the downloaded file. Pick the comma delimiter for US Excel and Google Sheets, or semicolon for the European locales where Excel expects it by default — Germany, France, the Netherlands, and much of Southern and Eastern Europe — and the columns line up immediately with no Text-to-Columns wizard required. Tab gives you a TSV you can paste straight into a sheet grid.

For Google Sheets, you can also copy the CSV output, open a blank sheet, and paste — Sheets detects the delimiter automatically and splits into columns. If you regularly pull JSON from an API and need it in Sheets, this browser tool is faster than writing a Google Apps Script for one-off jobs. No account needed, no quota to worry about, and the conversion finishes in under a second regardless of how many rows your JSON contains.

Choose your delimiter, headers, and columns

You control the output format. Toggle the header row on or off, switch the delimiter between comma, semicolon, and tab, and use the column chips to include or drop individual fields — handy when an API response returns thirty keys but you only need five. Column order always follows the order keys first appear in your JSON, so the CSV matches exactly what you expect without any alphabetical resorting.

The column chip interface also solves a common pain point: REST API responses often include server-internal fields like _id, __v, or updated_at that you do not want in a customer-facing spreadsheet. Click those chips to exclude them and the preview updates immediately. The selected column set is remembered for the current session, so if you paste updated JSON with the same schema it stays clean. For TSV output, the tab delimiter avoids any quoting conflicts with comma-heavy data like addresses or descriptions — a useful option when piping output to other command-line tools.

CSV to JSON with smart typing

Going the other way, paste or upload a CSV and the tool reads the first row as headers and every following row as an object, handling quoted fields with embedded commas and newlines along the way. The delimiter is auto-detected, and with Parse numbers & booleans on, 42 becomes a number and true becomes a boolean instead of strings, giving you clean, typed CSV to JSON output ready for an API or config file.

This is useful for importing spreadsheet data into a database, building a config file from a shared CSV, or feeding a frontend component that expects a JSON array. The output JSON is formatted with indentation for readability, not minified, so you can inspect it directly. If you need compact JSON for a production API payload, copy the output and minify it with any JSON formatter tool. The conversion handles up to thousands of rows without slowing down because it runs entirely in JavaScript in your browser tab, with no server round-trip adding latency.

How this compares to ConvertCSV, csvjson, and other online converters

ConvertCSV.com is a popular option but uploads your file to its servers, imposes a 1 MB file size limit on the free tier, and shows ads throughout the result page. csvjson.com is clean and developer-friendly but also server-side — your data crosses the network. json-csv.com offers a free tier but limits free conversions per day. Mr. Data Converter is minimal and browser-based but has not been maintained in years and lacks nested flattening.

This tool is fully client-side with no upload, no file-size cap (beyond device memory), no daily conversion limits, and no ads. For sensitive exports — customer records, API responses with personal data, internal analytics — nothing leaves your machine. You can disconnect from the internet after the page loads and keep converting offline. No sign-up, no cost, no compromise on privacy.

JSON vs CSV: hierarchical data versus flat tables

JSON (JavaScript Object Notation), standardised as RFC 8259 in 2017, was designed to represent hierarchical and nested data. An object can contain other objects or arrays to any depth: {"user": {"name": "Alice", "address": {"city": "NYC"}}} is valid JSON that carries structure naturally. This is why JSON became the dominant format for REST APIs, configuration files, and document databases — a single JSON document can encode a complete entity with all its relationships without any join or secondary table.

CSV (Comma-Separated Values), standardised as RFC 4180 in 2005, represents flat tabular data: rows and columns, exactly like a spreadsheet. There is no native concept of nesting. The same user record above cannot fit in a single CSV row without flattening — collapsing the nested address into a top-level column like address_city — or without encoding the nested object as a literal JSON string inside a cell. This fundamental structural difference is why conversion always involves a trade-off: JSON is better for APIs, configuration, and complex hierarchical data; CSV is better for spreadsheet software, bulk database imports, and data analysts who live in Excel or Google Sheets.

The practical implication is that JSON-to-CSV conversion is inherently lossy for anything beyond a flat array of objects. Once a deeply nested document is flattened to CSV, re-hydrating it back to the original structure requires knowing the original schema. For round-trip workflows — read from an API, inspect in a spreadsheet, write back to the API — keeping the JSON as the source of truth and using CSV only as an intermediate view is the safest pattern.

How JSON-to-CSV conversion works under the hood

Converting a JSON array of objects to CSV involves several non-trivial decisions. The first is header generation: because different objects in the same array may have different keys, a correct converter must take the union of all keys across all objects and use that union as the header row. An object missing a key simply gets an empty cell in that column — it is not an error. The second decision is key order: most converters follow first-seen order so the output column sequence is predictable and matches the first object in the array.

The harder problem is flattening nested keys. When an object contains a nested object, the converter must choose a separator for the composed key name: address.city (dot notation), address_city (underscore), or address-city (hyphen) are all common choices. Dot notation is the most widely understood but clashes with property names that already contain dots. Array fields introduce another branching decision: an array like "tags": ["node", "js"] can be handled three ways — stringified as a single cell value node,js, expanded into indexed columns tags.0 and tags.1, or expanded into multiple rows (one row per element, duplicating all other fields). Each approach suits different downstream consumers.

Type preservation is another structural issue. CSV is a string-only format: a JSON number 42 and a JSON string "42" are written identically in CSV. null becomes an empty string; undefined is also omitted as empty. true and false become the literal text true and false. When you convert CSV back to JSON without a type-inference step, every value is a string. The Parse numbers & booleans option in this tool re-applies type inference on the return trip, but if your data has a column that happens to contain only digits and is meant to be a string (a ZIP code like 07001, for example), you need to handle that explicitly — a leading zero in a ZIP code will be dropped if the value is parsed as a number.

CSV format edge cases and RFC 4180

RFC 4180 is a two-page memo, not a full standard, and it leaves enough unspecified to generate real interoperability problems. The most common gotcha is quoting: any field that contains a comma, a double quote, or a line break must be wrapped in double quotes. A double quote inside a quoted field must itself be doubled: the value He said "hello" is written as "He said ""hello""". Converters that do not handle this correctly produce malformed CSV that breaks when imported into Excel or a database. This tool wraps and escapes all such fields automatically on export and parses them correctly on import.

Line endings are another source of inconsistency. RFC 4180 specifies CRLF (\r\n) as the record separator. Many Unix-native tools emit only \n, and most modern parsers accept both — but older Excel versions on Windows expect CRLF and can produce an extra blank row at the end of a file when fed LF-only CSV. The BOM (Byte Order Mark) is a related issue: RFC 4180 does not mention encoding at all, but Excel on Windows historically opens CSV using the system locale encoding (often Windows-1252), which corrupts non-ASCII characters in names, addresses, and currency symbols. Adding the UTF-8 BOM (\xEF\xBB\xBF) as the very first three bytes signals to Excel that the file is UTF-8, preserving all Unicode content. The BOM is invisible in text editors and harmless to every parser that respects UTF-8, but it is the difference between a spreadsheet full of garbled characters and one that displays correctly.

Tab-separated values (TSV) sidestep the quoting-comma problem entirely: because a tab character almost never appears in real data, fields rarely need quoting at all, making TSV easier to generate and parse with simple string-split logic. TSV is common in bioinformatics, data science pipelines, and any workflow where the data itself contains many commas — postal addresses, product descriptions, and free-text fields are prime examples. The delimiter selector in this tool lets you output TSV with one click. One caveat: some TSV consumers do not handle quoted fields or embedded tabs, so treat TSV as the right choice when your data is clean and comma-heavy, not as a universal replacement for CSV.