Custom voices Paid

Custom voices let the bot read chat with a voice made from a recording you upload. Instead of a voice from the voice list, it reads with your own voice, or the voice of someone else who gave you permission.

Creating, turning on, sharing, importing and deleting are all done on the 'Membership' screen of the dashboard. The old Discord /membership customvoice command is gone.

Before you start

• Your personal membership has to be Plus or Premium to create or import a voice. → Features a membership unlocks
• You can keep up to 3 per account. Voices you created and voices you imported count together.
• You can only have one turned on at a time.
• Custom voices are tied to your account, not a server. If you turn one on, it's used on every server you use.
• Reading with a custom voice uses credits.

Where to do it

1. Log in to the dashboard with your Discord account.
2. Press 'Membership' in the sidebar (on a phone, the tab bar at the bottom).
3. Find the 'My custom voices' section under 'Status'.

With a Plus or Premium membership, this section appears even if you haven't made a voice yet. The number at the top right of the section (for example, 1/3) is how many voices you have now.

Creating a voice

1. Press [Create a voice]. The 'Create a custom voice' window opens.
2. In 'Name', enter a name you'll recognize in the list. Up to 20 characters.
3. Under 'Recording', press [Choose file] and pick a file.
4. Press [Create].

Item	Requirements
Format	mp3 · wav · m4a · ogg · flac · webm · mp4
Length	At least 3 seconds. Up to 30 seconds of speech is used
Size	Up to 50 MB
Name	Up to 20 characters. You can't use a name you already have

The window says "Upload a recording and we'll check and clean up noise and background sound, then create the voice. It can take a few minutes." When you upload the file, the window closes and a progress bar appears above the 'My custom voices' section.

Your recording is checked first

Your uploaded recording is checked first, and only the processing that recording needs is applied. The check looks at noise, background music, sound quality, echo, distorted sound and other people's voices. There's no separate option for choosing whether to improve sound quality. It's decided automatically from the check results.

Steps shown on the progress bar

The progress bar looks like this.

Creating 'My voice' · Removing background noise · done within a minute

Only the steps the check finds necessary are run. Not every recording goes through every step.

Progress bar text	What it does
Waiting for its turn	Waits while the voice of the person ahead is being made
Checking noise, background sound, and quality	Checks the condition of the recording
Removing background music	Removes music playing underneath
Boosting a quiet recording	Turns up sound that's too quiet
Reducing room echo	Reduces sound echoing in the room
Repairing clipped audio	Fixes sound that broke up because it was recorded too loud
Restoring sound lost to calls or compression	Fills in sound missing from call recordings and compressed files
Enhancing audio quality	Improves the overall sound quality
Removing background noise	Removes surrounding noise
Reducing light noise	For clean recordings, only reduces small noise
Trimming pauses and evening out the volume	As a final step, evens out pauses and volume
Picking 30 seconds of speech from 2m 10s	If the recording is long, picks 30 seconds of speech to use
Registering the voice	Makes the voice from the cleaned-up recording
Waiting for registration	Waits briefly because many voices are being registered

If people are waiting, something like '2 ahead of you' is added at the end. If it can tell when it'll finish, something like 'done in about 3 min' is added.

• It can take a few minutes. It keeps going even if you go to another screen. When it's done, it appears in the list.
• While a voice is being made, you can't create or import another one. You can again when it's done.

Check results

After the recording is checked, the results appear below the progress bar.

Noise some · Music None · Quality Moderate
Added: Volume boost · Noise removal

Line	What it shows
Noise	almost none · some · a lot
Music	yes · None
Quality	good · Moderate · low. If there's distorted sound, '(distorted)' is added
Added	Whichever of these were applied: Background music removal · Volume boost · Noise removal · Light noise reduction · Echo reduction · Distortion repair · Quality restoration

If no processing is needed, 'Clean recording, so it only gets trimmed' appears.

Alerts

If the check finds something to watch out for in the recording, an alert appears below the results. Even when an alert appears, the voice is still made.

Alert	Next time you record
Another person's voice is mixed in.	Please listen to the result.
There was a lot of echo, so it was reduced.	Recording in a small room will sound more like you.
Parts are distorted because it was too loud.	Recording a little farther from the mic sounds cleaner.
The recording was quiet, so it was turned up and cleaned.	Recording closer to the mic sounds cleaner.

When it's done

The progress bar changes to this.

Created 'My voice'. Press 'Use this' below to start using it.
Made from 21s of speech

If the recording was long, it shows something like 'Made from 30s of speech out of 2m 10s'. The check results and alerts stay there even after the voice is made.

If it couldn't be made, 'Couldn't create' is followed by the reason and what to do. You can clear the progress bar with [Close].

Recording tips

• A recording of one person speaking clearly works best. If other voices are mixed in, the alert 'Another person's voice is mixed in.' appears.
• About 30 seconds is enough. Anything longer isn't used.
• Noise and background music are removed automatically, but you'll get a better result if you record somewhere quiet to begin with.
• Only upload another person's voice if they've given you permission.

What you can do in the list

Each voice row has a name, the date it was created (or imported), labels and buttons.

Label	Meaning
In use	This voice is reading now
Imported	You imported a voice someone else shared
Please check	The voice's status needs to be checked. Please contact us on the official server

Button	What it does
[Use this]	Reads with this voice. Any other voice you had on is turned off
[Turn off]	Turns off custom voices and reads with the default voice
[Share]	Shows the share code
[Delete]	Deletes the voice. You're asked once more

When you turn a voice on or off, the change applies from the next sentence. If no voice is turned on, the bot reads with the voice you picked in Choosing a voice.

Sharing

When you press [Share], a share code like ABCD-EFGH-JKLM-NPQR appears under the row. Press [Copy] to copy the code. Send this code to the other person.

• Anyone who knows the code can use this voice with 'Import with a code'.
• The same voice always shows the same code, whenever you press it.
• Imported voices can't be shared again. Ask the person who created it for the code.
• If you delete the voice, it can't be imported with that code anymore. People who already imported it can keep using it.
• Share codes in the same format that you got from the Discord bot before still work.

Press [Hide code] to fold the code away.

Importing with a code

1. Press [Import with a code].
2. Enter the code you received in 'Share code'. You can type it in lowercase or leave out the hyphens.
3. In 'Name (optional)', enter the name to use in your list. If you leave it empty, it's saved with its original name. Up to 30 characters.
4. Press [Import voice].

After importing, you can press 'Use this' to start using it right away. Imported voices count toward your 3 too.

If you see this	What to try
Couldn't find that share code. Please check the code and try again.	Check the code again. The creator may have deleted the voice
You've already imported this voice.	It's already in your list
It's your own voice, so it's already in your list.	It's the code for a voice you created
You already have a voice named '○○'. Please enter a different name.	Enter a different name in 'Name (optional)'
You can keep up to 3. Delete one to import another.	Delete a voice you don't use

Deleting

When you press [Delete], you're asked 'Delete '○○'?'. Read the text below, then press [Delete] or [Cancel].

Situation	Confirmation text	When you delete it
A voice you created that no one has imported	This can't be undone.	The voice is gone
A voice you created that others have imported	This can't be undone. The ○ people who imported this voice can keep using it.	It's removed from your list, and people who imported it keep using it
An imported voice	It's only removed from your list. To use it again, import it with the share code again.	It's removed from your list only

CAUTION

If you delete a voice you created, it can't be undone. To use the same voice again, you have to create it again from a recording. A newly created voice also gets a different share code, so it can't be imported with the old code.

Rarely, you may see 'Couldn't fully delete '○○', so it's turned off for now. Please delete it again shortly.' The voice stays in the list, turned off, so press [Delete] again a little later.

When your custom voice isn't used

If a custom voice can't be used, the bot reads with another voice without telling you. Check the items below from the top.

What to check	How to check
Did you turn the voice on?	See if it has the 'In use' label in the list
Can it be used right now?	See if the notice 'Custom voices can't be used right now' is above the list
Has the server turned on 'Lock voice'?	On servers with 'Lock voice' on, custom voices aren't used → My settings
Is there a 'Please check' label?	Please contact us on the official server

Even if you can't use it right now because of your membership or credits, the voices you made aren't deleted. When you can use them again, they'll be just as they were.

Related pages

• Features a membership unlocks
• Membership status and storage
• Choosing a voice
• My settings
• How the bot reads