Voice Inputs

Speak naturally to Playlab and transcribe text in real-time with voice input

🧪 This feature is currently in BETA and is currently available to Organizations only.

What is this feature?

Voice Inputs enables real-time speech-to-text transcription directly within the Playlab interface, providing a multimodal interface for natural language interaction. The feature leverages speech recognition to convert spoken input into text with high accuracy, supporting a more accessible and efficient workflow for users across different contexts and abilities.

This implementation provides an alternative input modality to provide greater accessiblity and to support those who simply prefer verbal communication. Voice and text input can be used interchangeably or in combination.

Voice inputs require microphone permissions in your browser. You’ll be prompted to grant access when you first use the feature.

What is the Rationale for this feature?

Typing can be slow and cumbersome, especially when you’re working through complex ideas, brainstorming, or need to communicate detailed instructions. Many people can speak much faster than they can type, and speaking feels more natural for certain types of communication. This creates friction in workflows where you want to quickly capture thoughts or have fluid conversations with AI.

The Voice Inputs feature eliminates these barriers by providing speech-to-text transcription directly in Playlab. Instead of switching between typing and speaking, or using external dictation tools, you can now speak your thoughts directly into any Playlab conversation. The system handles the transcription in real-time, providing immediate visual feedback so you know exactly what’s being captured. This makes it easier to work with Playlab in a more natural, conversational way while maintaining your flow of thought.

How to Use Voice Inputs

1. Enable Voice Input at Organization Level

   Your organizational owner needs to enable voice input in organization settings by toggling on “Allow Voice Inputs to Apps”.

   We do not recommend enabling this feature for organizations with students under the age of 13.

   
2. Enable Voice Input in App Builder

   In the app builder settings, toggle on “Allow Voice Input” to enable voice input functionality for your specific app.

   
3. Locate the Microphone Icon

   Look for the microphone icon in the chat input bar at the bottom of your Playlab interface.

   
4. Grant Microphone Permission

   Click the microphone icon. If this is your first time using the feature, your browser will prompt you to grant microphone access to Playlab. Click “Allow” to proceed.

5. Start Speaking

   Once the microphone icon turns blue with a glowing ring animation, you’re actively recording. Simply start speaking naturally, the transcription will appear in real-time in the chat input field. Right now, voice input supports English, Spanish, French, German, Italian, and Portuguese. We are looking to add more language support as we expand this feature.

   
6. Monitor Your Recording

   Watch the visual indicator to confirm you’re recording. The glowing ring animation provides clear feedback that the system is capturing your voice. A countdown timer will appear as you approach the recording time limit.

7. Stop Recording

   Click the microphone icon again to stop recording. Your transcribed text will remain in the chat input field, ready for you to review or edit.

8. Review and Send

   Review the transcribed text for accuracy. You can edit it manually if needed, add additional context by typing, or click send to submit your message to Playlab.

Best Practices

Speak Clearly and Naturally

The transcription system works best when you speak at a normal conversational pace. You don’t need to speak slowly or enunciate excessively, just speak as you naturally would in a conversation. The system is designed to handle natural speech patterns.

Use in Quiet Environments

For best results, use voice input in relatively quiet environments. Background noise, music, or other conversations can interfere with transcription accuracy. If you’re in a noisy environment, consider moving to a quieter space or using typed input instead.

Stick to One Language Per Recording

The transcription system performs best when you speak consistently in one language throughout a recording session. Switching between languages mid-recording may reduce accuracy. If you need to communicate in multiple languages, consider using separate recordings for each language.

Be Mindful of Recording Limits

Voice recordings have a maximum time limit to prevent accidental extended recordings. A countdown timer will appear during the final seconds before the recording automatically stops. Plan to break longer messages into multiple recordings if needed.

Combine Voice with Text Input

You can use voice input alongside traditional typing. If you have some text already in the chat input, clicking the microphone will preserve that text and append your voice transcription. This makes it easy to mix both input methods as needed.

Review Before Sending

While the transcription is highly accurate, it’s always good practice to quickly review the transcribed text before sending. This helps catch any transcription errors or unclear phrasing that might need clarification.

Known Issues

As this feature is currently in beta, there are a few known limitations to be aware of:

First Few Words May Be Missed

Issue: There can be a brief initialization period when you first click the microphone button. In some cases, if you start speaking immediately, the first few words might not be captured.

Workaround: Wait for the microphone icon to turn blue with the glowing ring animation before you start speaking. This visual indicator confirms that transcription is active and ready to capture your voice.

Cannot Type While Recording

Issue: To prevent conflicts and ensure clean transcription, you cannot type into the chat input field while actively recording with the microphone.

Workaround: This is intentional behavior. Simply stop your recording by clicking the microphone icon again, then you can type additional text or edits as needed.

Language Switching Challenges

Issue: If you start speaking in one language and then switch to another language mid-recording, the transcription accuracy may decrease.

Workaround: Use consistent language within each recording session. If you need to communicate in multiple languages, stop the current recording and start a new one for the different language.

Privacy & Security

Your privacy and security are important to us. Here’s what you should know about voice input:

• Microphone Access: Voice input requires microphone permissions in your browser. You’ll be prompted to grant access when you first use the feature.
• Data Processing: This feature uses AssemblyAI to enable users to transcribe text responses using their voice. Playlab and AssemblyAI will never use voice recordings to train Playlab.
• Under 13 Use: We recommend disabling this feature if your organization has students under age 13 using Playlab.
• Control: You have complete control over when the microphone is active. The visual indicators (blue microphone icon and glowing ring) always show when recording is in progress.
• Secure: All voice data transmission uses secure protocols to protect your privacy.

Frequently Asked Questions

How accurate is the voice transcription?

The voice transcription system is highly accurate for clear speech in supported languages, particularly English. Accuracy depends on factors like audio quality, speaking clarity, background noise, and accent. The system is designed to handle natural conversational speech patterns effectively.

What languages are supported?

Voice input works particularly well for English. Support for additional languages may vary in accuracy. The system performs best when you speak consistently in one language throughout a recording session.

How long can I record for?

Voice recordings have a maximum time limit (typically at least one minute) to prevent accidental extended recordings. A countdown timer will appear during the final seconds (usually the last 10 seconds) before the recording automatically stops. If you need to communicate more, you can start additional recordings.

Can I edit the transcription after recording?

Yes! Once your voice is transcribed into text, it appears in the chat input field just like any typed text. You can edit, add to, or modify the transcribed content before sending your message.

What if the transcription makes a mistake?

Simply edit the text in the chat input field to correct any transcription errors before sending your message. You can manually type corrections or use voice input again to re-record sections that weren’t captured accurately.

Can I use voice input on mobile devices?

Yes, voice input is available on mobile devices through the Playlab mobile app. The feature works the same way as on desktop, using your device’s microphone.

Does voice input work with all Playlab apps?

Yes, voice input is a core feature of the Playlab chat interface and works across all Playlab apps and conversations. Any app that accepts text input can be used with voice input.

What happens if I walk away during recording?

The recording will automatically stop when the maximum time limit is reached. The countdown timer in the final seconds provides a warning. This prevents accidentally leaving the microphone on indefinitely.

Can I use voice input if I already have text in the input field?

Yes! If you have text already in the chat input, clicking the microphone will preserve that existing text and append your voice transcription to it. This makes it easy to combine typed and spoken input.

Is voice output or text-to-speech available?

Yes! Voice Output is now available in Playlab. You can listen to Playlab’s responses using text-to-speech playback. Check out the [Voice Output documentation](/features/Voice Output) to learn more about how to use this feature.

We Want Your Feedback!

Thank you for trying out Voice Inputs in Playlab. We’re continuously working to improve the accuracy, reliability, and user experience of this feature. Since this is still in beta, your feedback is invaluable.

Contact us at [e-mail]

---

Last updated: March 15, 2026