RELATIONAL MODEL
Schema overview
0tables
0relations
0rows
Tablecraft converts nested JSON into a relational model. Each array becomes its own table with a primary key and a foreign key back to its parent, so you can see the structure at a glance and export a SQL schema. It also reads HTML tables. Everything runs in your browser.
Paste JSON, HTML or CSV into the editor, or drop a .json, .html or .csv file.
Press Build tables (or Ctrl or Cmd+Enter). Nested arrays become child tables.
Switch between the Relationships view and the Data preview to check the result.
Download the SQL (schema, or schema with INSERT statements) or any table as CSV.
Arrays of objects become tables linked by foreign keys.
Reads <table> elements and infers keys from columns such as customer_id.
Cards show each table, its columns and types, with PK and FK marks.
Click any cell to fix it, add a row, and every export uses your changes.
Transpose, delete empty rows, remove duplicates, change case, find and replace with regex, and undo or redo.
Read CSV or TSV, then export CSV, Excel, Markdown, HTML, JSON, XML, YAML or LaTeX.
Filter rows across every table before you export.
CREATE TABLE with inferred types, plus optional INSERT statements for every row.
Your data never leaves your browser.
Download SQL, an Excel workbook, CSV or JSON, or copy Markdown or HTML.
Yes. It is free with no sign-up.
No. The conversion runs in your browser.
Generic CREATE TABLE statements with quoted names, primary keys and foreign keys. Tick "Include data" in the Data preview to add INSERT statements for every row.
Tables are named from the JSON keys. Child tables get a foreign key named after the parent, such as customer_id, linked to the parent's id.
A key is generated for each row so the child tables can still link to it.
It is limited by your browser's memory. Very large files may be slow.