This article explains how to change what the AI reads aloud in voice narration. There are two ways to do it.

Also, edits are shared across the entire team and are reflected in the readings of other SOPs as well.

Registered readings are also used in search. When you search with hiragana, SOPs containing that word will show up in the results (Full-text search feature).

① Edit while viewing an SOP

1. Click the [Report misreading] icon
   
2. Select the relevant part (in this example, MΩ)
   

3. Enter the correct reading and click [Register]. When you open it again, the text will change to what you specified.
   

   
   Note: By default, changes made by a Group Administrator or less become a "Report," and they are applied once a Team Administrator or higher checks the contents and approves them.

② Configure in advance from Dictionary / Templates

You can also open [Dictionary / Templates] from the sidebar and configure readings there.

③ Bulk import from Excel/CSV

To register or update a large number of readings, you can import an Excel/CSV file from the [Bulk import] button on the [Pronunciation Dictionary] tab. This is handy when you want to apply a glossary or an in-house list of readings all at once.

1. Open [Dictionary / Templates] in the sidebar and switch to the [Pronunciation Dictionary] tab
2. Click [Bulk import] in the upper right
3. Click [Excel template] in the upper right to download the template. If readings are already registered, the template is downloaded with those readings already filled in, so you can edit it as-is and upload it again
4. Drag and drop the template (or your own CSV / Excel file with the columns "Written form / Reading / Language (Optional)"), or paste the contents
5. Difference preview appears, and you can check the contents in four categories: Add/Update/Skip/Error
6. If there are no issues, click [Register N items] to confirm

• Supported file formats: CSV, TSV, Excel (.xlsx / .xlsm). Character encoding (UTF-8 and Shift-JIS) is detected automatically
• Language column: If omitted, entries are registered as Japanese (ja)
• Japanese readings must be entered in hiragana or katakana. Other characters (such as kanji only, English letters only, or symbols only) will cause an error
• Readings that are Same as existing are skipped automatically
• After registration, applying the changes to existing audio (regenerating it) may take a few minutes
• Even when a user at Group Administrator or less runs a bulk import, just as with a regular addition, a Report requesting approval is sent to a Team Administrator or higher

Related dictionary features

The following dictionary is also available in [Dictionary / Templates]. It also supports bulk import from Excel/CSV.

• Terminology Glossary: Specify in advance the translations used for automatic translation, and alternative names to unify variations in notation (see: Terminology Glossary)