CSV to SQL

Table name

Dialect

Include CREATE TABLEEmpty cell becomes NULLOne multi-row INSERT

SQL Output

SQL statements will appear here...

How to Convert CSV to SQL Online

Paste your CSV, or upload a .csv file

The first row is read as the header and becomes the column list.
Type the target table name and pick the dialect

PostgreSQL, MySQL, or SQLite.
Adjust the options if you need to

turn off CREATE TABLE when the table already exists, keep empty cells as empty strings instead of NULL, or switch to one multi-row INSERT.
Copy the SQL, or download it as a .sql file and run it against your database yourself

Common Use Cases

Seeding a development or test database from a spreadsheet export

Moving a report out of Excel and into a real table so you can query it

Turning a data dump from a third-party service into rows you can import

Writing a migration fixture without hand-typing a hundred INSERT lines

Why Use Our CSV to SQL Converter?

Loading a spreadsheet into a database usually means either writing the INSERT statements by hand or uploading the file to a site you do not control. Neither is appealing when the CSV holds customer records or unreleased figures. This converter reads the CSV in your browser, infers a type for each column, and writes the SQL you can paste straight into psql, the MySQL client, or the SQLite shell. It never connects to a database, and on this page the file is not uploaded.

CSV is how data leaves almost every system: an export button, a report download, a share from a colleague. SQL is how it gets into a database. The gap between the two is usually crossed by hand, one INSERT at a time, and that is exactly the kind of repetitive work worth removing.

On this page, the converter reads the CSV in your browser. The REST API and the MCP tool are a separate path: a call sends the CSV you include to FindUtils, which converts it without storing it. The converter splits the header row into a column list, samples the first fifty data rows to infer whether each column is an integer, a decimal, a boolean, or text, and then writes a CREATE TABLE followed by the INSERT statements. Quoting follows the dialect you choose, so the output runs without editing whether you paste it into PostgreSQL, MySQL, or SQLite.

Nothing is executed. The tool produces SQL text and hands it to you; connecting to a database and running it stays your decision, on your machine, with your credentials.

How it compares

Most online CSV to SQL sites upload the file to a server to do the parsing. For a public dataset that is fine. For an export of customer emails, a payroll sheet, or anything under an NDA, it means handing the file to a third party for a job a browser can do on its own. On this page, this converter makes no such request.

CSV to SQL Tips

Types are inferred from the first 50 data rows. A column that mixes 1 and 2.5 becomes numeric; a column with one non-numeric value becomes text.
An empty cell never narrows a type. A column of integers with a gap is still an integer column.
Identifiers are quoted for the dialect you pick: backticks for MySQL, double quotes for PostgreSQL and SQLite.
A value containing a single quote is escaped by doubling it, which is portable across all three databases.
SQLite has no boolean literal, so true and false are written as 1 and 0 for that dialect only.

Frequently Asked Questions

Does my CSV get uploaded anywhere?

Not on this page. The CSV is parsed in your browser and the SQL is generated there too. Open your browser developer tools, switch to the Network tab, and convert a file: you will see no request carrying the CSV. That makes the tool safe for exports containing customer records, financial figures, or anything else you cannot share with a third party. The API is a separate path: a call sends the input you include to FindUtils over HTTPS, where it is processed and the result returned. FindUtils does not store that input, and logs the request without its content.

How are column types decided?

Each column is sampled across the first 50 data rows. A column whose values are all whole numbers becomes an integer type; one that mixes whole numbers and decimals becomes numeric; one containing only true and false becomes boolean; anything else becomes text. Empty cells are ignored during inference, so a column of integers with a few gaps is still an integer column. If a single value does not fit, the column widens to text rather than producing SQL that fails on import.

What happens to a comma inside a quoted value?

It is preserved. The parser follows RFC 4180: a value wrapped in double quotes may contain commas, and a doubled quote inside such a value means one literal quote. A cell reading "Doe, John" becomes the single SQL literal 'Doe, John' rather than two columns.

How are empty cells handled?

By default an empty cell becomes NULL, which is almost always what you want in a database. Turn off "Empty cell becomes NULL" to write an empty string instead, which matters when a column is NOT NULL or when an empty string and a missing value mean different things in your schema.

Which SQL dialects are supported?

PostgreSQL, MySQL, and SQLite. The dialect changes two things: how identifiers are quoted, and which type names appear in the CREATE TABLE. MySQL uses backticks around names; PostgreSQL and SQLite use double quotes. Type names follow each database's conventions, and SQLite stores booleans as 1 and 0 because it has no boolean literal.

Can I skip the CREATE TABLE?

Yes. Turn off "Include CREATE TABLE" and the output is INSERT statements only, which is what you want when the table already exists and you are just loading rows into it.

What is the difference between one INSERT per row and a multi-row INSERT?

One statement per row is easier to read and easier to run partially: if one row fails, the others still land. A single multi-row INSERT is considerably faster for large imports because the database plans it once, but it is all-or-nothing inside a transaction. Use per-row for a few dozen rows and the multi-row form for thousands.

Does it handle a CSV with a semicolon or tab delimiter?

Not yet — the parser reads comma-separated values. If your export uses semicolons or tabs, run it through the CSV to JSON side of the JSON CSV Converter first, or open it in a spreadsheet and re-export with commas.

What happens to a cell containing JSON?

It is written as a quoted SQL string containing the JSON text, with its internal quotes escaped. It is never rendered as the literal text [object Object], which is a common failure in converters that stringify without checking the value type. If the target column is a JSON or JSONB column, the string literal will cast cleanly on insert.

Is there a row limit?

On this page, the conversion runs in your browser, so the limit is your browser's memory rather than a server setting. Tens of thousands of rows convert without trouble on a desktop machine; very large files are better handled by your database's own bulk loader, such as COPY in PostgreSQL or LOAD DATA INFILE in MySQL.

Run it from code

$ curl -X POST \
  api.findutils.com/api/tools/csv-to-sql/execute \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"csv":"id,name\n1,Ada\n2,\"Doe, John\"","table":"people","dialect":"postgres"}'

$ claude mcp add findutils \
  --transport http https://mcp.findutils.com/ \
  --header "Authorization: Bearer YOUR_API_KEY"
# then call findutils:csv_to_sql

Running a tool needs an API key. Keys are not for sale yet: ask for early access. About API keys

The REST API and the MCP tool are a separate path from this page: a call sends the input you include to FindUtils over HTTPS, where it is processed and the result returned. FindUtils does not store that input, and logs the request without its content.

API reference Add to Claude