Upload a JSON file and read it as a document: a list of records becomes a table, an object becomes sections with its fields above them.

Upload a .json file. A list of records becomes a table; nested objects become headings. One value per line — a log export — is understood too.

What comes across

JSON has no headings or paragraphs, so the converter decides by shape. The file's name, without its extension, becomes the document's heading: users.json opens as # users.

In the file	In the Markdown
An array of objects whose values are all plain	A table, one row per object, keys as columns
An array of plain values	A bullet list
An object	Its plain fields as bold labels (`Status: shipped`), then a heading per nested field (`## Customer`)
An array of mixed or nested items	A numbered section per item: `## 1`, `## 2`
An object or mixed list three levels below the title	A fenced `json` block, pretty-printed
`null`	`_null_`, written rather than skipped
An empty array or object	`_Empty list._` or `_Empty object._`

Columns are collected from every record in the order they first appear, so a field only the tenth record has still gets one, and the others get an empty cell. Keys become words: underscores and hyphens turn into spaces, camelCase is split and the first letter capitalised, so first_name reads "First name" and lastLogin "Last Login".

What is left behind

The keys as written. Outside a fenced block, only the reworded form appears.
Number formatting. Numbers are read as numbers, so 42.50 comes out as 42.5, 1e3 as 1000, and an integer longer than about 15 digits can lose its last ones. Values in quotes are untouched: "007" stays 007.
The table, if one record is nested. One record holding a list, even an empty one, turns the whole array into numbered sections. Nothing is flattened into dotted column names.
Escaping. Strings are written as they are, so asterisks or underscores in a value are read as Markdown. Inside a table, a pipe is escaped and a line break becomes <br>.

An example

A list of users from an API, saved as users.json:

[
  { "id": 1, "first_name": "Ana", "isActive": true, "lastLogin": "2026-10-01" },
  { "id": 2, "first_name": "Ben", "isActive": false, "lastLogin": null },
  { "id": 3, "first_name": "Chloé", "isActive": true, "team": "Docs | Web" }
]

It becomes a # users heading and this table:

| Id | First name | Is Active | Last Login | Team |
| --- | --- | --- | --- | --- |
| 1 | Ana | true | 2026-10-01 |  |
| 2 | Ben | false | _null_ |  |
| 3 | Chloé | true |  | Docs \| Web |

Limits

A file can be up to 10 MB, and the page takes .json only, so a .jsonl or .ndjson file has to be renamed first. You can also paste the text.
The file is converted in your browser and is not uploaded. If you are signed in, the document reaches your account only when you press Save, and a saved document can be at most 4 MB.
The JSON has to be strict. A file that will not parse is refused with "That is not valid JSON" and the browser's own message, which usually names the position.

Questions

How do I convert JSON to a Markdown table?

An array of objects whose values are all strings, numbers, booleans or null becomes one table. Flatten nested records first, with jq for instance, or accept sections instead. JSON to a Markdown table covers what converts cleanly and what does not.

What happens to nested JSON?

Each nested field gets a heading, down to three levels below the title. At that level, an object or mixed list becomes a fenced json block, while a table or plain list is still drawn.

Can I convert JSON to Markdown in Python?

There is no Python package; the page converts in the browser. With an account, the API runs the same code: send the file to POST /api/v1/documents?kind=json-to-markdown&name=users.json with an API key, using requests or any HTTP client, and the result is saved in your account, headed # users.

Does it read JSON Lines?

Yes, once the file ends in .json. When the whole file does not parse, each non-empty line is read as a value of its own, so a log of one flat object per line becomes a table. If any line fails, the error from the first attempt is reported.

Why does it say my file is not valid JSON?

The usual causes are a trailing comma after the last item, single quotes instead of double, and comments, which JSON does not allow. How to open a .json file explains how to find the fault from the position in the message.

Can it turn Markdown back into JSON?

No. The downloads are Markdown, HTML, Word, plain text, EPUB and formats for Confluence, Jira and Slack; JSON is not among them.