Import a CSV word list with weights

1. Save a CSV with word and weight columns

Save your spreadsheet as a comma-separated UTF-8 CSV. The first line must be word,weight, with exactly those two columns in that order. The importer accepts LF or CRLF line endings and a UTF-8 byte-order mark. If your spreadsheet uses semicolons, choose a comma-separated export before importing.

word,weight
calm,2
calm,3
bright,1

The word column can contain a single word or a short phrase. Spaces within a phrase stay together; spaces at the beginning and end are trimmed. Text is normalized to Unicode NFC to handle equivalent accented characters consistently. Each weight must be a number greater than zero. Values such as high, Infinity or NaN won't work.

2. Put commas and quotation marks inside quoted fields

If a phrase includes a comma, surround the whole field with double quotation marks. To include a quotation mark within that field, write it twice, as in the example below. Line breaks and control characters aren't allowed within a word or phrase, even if the surrounding CSV quotes are correct.

word,weight
"Clear goals, shared notes",4
"Say ""thank you""",2

These rows become Clear goals, shared notes and Say "thank you". The CSV quotes tell the importer where the field begins and ends; the doubled marks become visible quotation marks in the phrase. Check the wording in the imported table so the punctuation appears exactly as you intend.

3. Import the file and fix any errors

Open the CSV word cloud generator and choose Import TXT / CSV. The file is read in your browser. If a data row is invalid, the import stops and shows the source line so you can fix it. The importer won't bring in only the rows it could read and leave the rest out.

word,weight
calm,2
bright,0
"clear goals, shared notes",4
focus,three
crowded,1,extra

A broken quotation mark can stop the file from being parsed before later rows are checked. Fix the first error and import again. If you don't know the correct weight, check the source instead of choosing a number just to clear the error. For decorative lists, you can choose weights deliberately to control emphasis.

4. Choose whether to merge duplicates

In the first example, merging duplicates combines the two calm rows into calm at weight 5, alongside bright at weight 1. With merging off, all three rows remain separate. Both versions have a total weight of 6. Merging compares the normalized text after your case setting; it doesn't combine synonyms or remove accents.

CSV import preserves uppercase and lowercase by default. If you turn on Convert to lowercase, check the resulting words before generating. Accents are kept, so café and cafe remain different. The parsing locale controls case handling for new input; it doesn't change the language of the website or rewrite saved words.

5. Check the list size

A CSV can contain up to 2,000 data rows and must be no larger than 512 KiB, or 524,288 bytes. The resulting editable list can contain up to 1,000 rows, with no more than 100 Unicode characters in each word or phrase. Keeping duplicates separate uses a row for each one. Files over these limits are rejected.

You can arrange up to 250 selected rows in one layout. For a longer list, choose how many top-weighted words to include and check the report for the rest. A successful CSV import confirms that the list is valid; it doesn't guarantee that every selected phrase will fit the chosen shape.

6. Save the list and finished project

Generate the cloud and check the placement report. Download project JSON to keep the words, settings and layout, or Word list CSV for the words and weights alone. The CSV exporter protects text that a spreadsheet might otherwise treat as a formula. Check the downloaded file in the app where you plan to use it.