Voicely Translate

Voicely Translate is a Discord bot that automatically transcribes and translates conversations happening in voice channels.

When the bot is active in a voice channel, it listens to each speaker individually, transcribes what they say, and posts the original transcription alongside translations in the voice channel's side chat.

Multiple translation languages can be enabled at once, and languages can be added or removed while the bot is already translating.

Features

• Automatically transcribes speech from Discord voice channels.
• Translates conversations into multiple languages at the same time.
• Posts transcriptions and translations directly in the voice channel's side chat.
• Supports multiple people speaking, including overlapping speakers.
• Lets you add or remove translation languages without restarting the session.
• Uses BCP 47 language tags, allowing you to request languages beyond a fixed built-in list.
• Automatically leaves an empty voice channel after a configurable timeout.
• Can optionally be paired with Voicely Text to read translated messages aloud.

Starting a Translation Session

Use:

/join

The languages option is optional. If you specify languages, those languages will be used for the translation session:

/join languages:<language tags>

Separate multiple languages with commas.

For example:

/join languages:en,ja

This starts a translation session with English and Japanese enabled.

You can enable as many languages as you need:

/join languages:en,ja,es,fr,de,ko

If you use /join without specifying any languages, Voicely Translate will use the server's default translation languages.

If the server does not have any default languages configured, the bot will ask you to specify languages or ask a server administrator to configure the defaults.

Once connected, Voicely Translate will post its transcriptions and translations in the side chat.

Language Tags

Voicely Translate uses language tags rather than language names.

For example:

• en — English
• ja — Japanese
• es — Spanish
• fr — French
• de — German
• ko — Korean
• zh — Chinese
• haw — Hawaiian

Regional and script-specific BCP 47 tags can also be used, such as pt-BR or zh-TW.

The bot is not limited to the languages shown by /languages. You can also try other valid BCP 47 language tags.

Voicely Credits

Voicely Translate uses a credit system to pay for the transcription and translation services required to process voice conversations.

100 Voicely Credits = $1.00 USD.

New servers currently receive 50 free trial credits. That is enough for approximately 1000 5-second voice messages, although actual usage varies.

Credits are consumed as speech is transcribed and translated. The amount used can vary depending on how much people speak and how many translations are requested.

Enabling more translation languages can increase usage because each transcription needs to be translated into the requested languages.

Estimated Monthly Usage

The number of credits a server uses depends on how much people speak and how many translation languages are enabled. The table below provides approximate monthly usage based on typical 5-second voice messages. Actual usage may vary.

How Translation Works

When someone speaks in the voice channel, Voicely Translate processes that person's speech independently.

The basic process is:

1. Voicely Translate receives the speaker's voice audio.
2. The audio is transcribed into the language actually being spoken.
3. The transcription is translated into the currently enabled languages.
4. The original transcription and translations are posted in the voice channel's side chat.

Because speakers are processed independently, the bot is designed to handle conversations where multiple people may speak at or near the same time.

Short pauses are used to determine when an utterance has finished before it is sent for transcription.

Nonverbal sounds and standalone hesitation noises may be ignored rather than posted as messages.

Example

Imagine three people are in a voice channel and English, Japanese, and Spanish are enabled.

Someone says:

Are we ready to go?

Voicely Translate can post the original transcription along with translations such as:

Original — English Are we ready to go?

Japanese もう行く準備はできた？

Spanish ¿Estamos listos para irnos?

The exact formatting and translations may vary.

Command Reference

Using Voicely Text

Voicely Translate is designed to display translations as text.

If you also want translated messages to be read aloud in the voice channel, you can use Voicely Translate alongside Voicely Text.

When Voicely Translate joins a server that does not have Voicely Text installed, it may provide a link for adding it.

Tips for Better Results

For the best transcription and translation quality:

• Speak clearly and at a normal volume.
• Avoid extremely loud background music or noise.
• Use the most appropriate language tags for the conversation.
• Remember that automated transcription and translation can make mistakes, especially with names, slang, very short utterances, heavy background noise, or ambiguous speech.
• Use regional language tags when the distinction is important.

About Voicely Translate

Voicely Translate is designed to make multilingual Discord voice conversations easier by putting the original speech and translations together in the same place where the conversation is happening.

Instead of requiring everyone in a call to speak the same language, participants can speak naturally while using the voice channel's side chat to follow the conversation in the languages they understand.

Support

If you have any problems with the bot or want to request a feature, please create an issue, and I will address it to the best of my ability!