Feature Request Priority: Text-to-Speech for AI Assist / Chatbot Replies

Hello tawk.to Support Team,

I’m reaching out to strongly request that text-to-speech functionality be prioritized for AI Assist and chatbot conversations.

I saw this feature discussed in the community feature requests, including the request for:

“Speech-to-text — guests talk to the bot
Text-to-speech — bot reads the reply out loud”

While both would be valuable, I believe text-to-speech (bot reads the reply out loud) should be treated as the immediate priority. Honestly both.

For businesses using AI Assist seriously, the chatbot is no longer just a text box. It is becoming a digital representative of the brand. In my case, it is carefully built with a strong knowledge base, detailed FAQs, and a very specific communication style. The way the bot responds is a major part of the customer experience. But when the response is only text, a large part of that experience is lost.

Even a simple speaker button beside each AI response would make a major difference. The customer could tap the button and hear the reply read out loud. It does not need to be fully conversational or bi-directional at first. A basic text-to-speech option with a small selection of male and female voices would already be a powerful improvement.

This would help with:

Accessibility for users who have difficulty reading or prefer audio
Mobile users who may not want to read long replies on a small screen
Customer engagement, because voice feels more natural and interactive
Brand experience, especially for companies that put serious work into their AI assistant’s tone and personality.
Conversion, because customers are more likely to stay engaged when the assistant feels alive and responsive.

This is not just a cosmetic feature. Voice is becoming a core part of how people expect to interact with AI. Many users are already comfortable with assistants that speak back, and chat platforms that do not support this will start to feel limited compared to newer AI experiences.

I believe this is something tawk.to can figure out and implement easily in a practical way. Even if it starts simple, the feature would immediately make AI Assist feel more modern, accessible, and valuable.

Please consider moving text-to-speech higher on the development roadmap. For businesses like mine, this would be a major upgrade and could make the chatbot experience much stronger for customers.

Thank you,

Richard Barton | Skan Digital
Founder & CEO
DigiDevice

Hi iamDigital,

Thank you for taking the time to share such detailed and thoughtful feedback regarding text-to-speech and speech capabilities for AI Assist.

We truly appreciate the effort you put into explaining the practical business value, accessibility benefits, and overall customer experience improvements this feature could provide. Your points about AI Assist becoming a digital representative of a brand and the importance of voice interaction in modern AI experiences are very insightful.

We’ll make sure your feedback and feature request are raised with the appropriate team for consideration in future development discussions. While we’re unable to provide a timeline or guarantee implementation at this time, detailed use cases and business perspectives like yours are extremely valuable in helping the team evaluate potential improvements.

Thank you again for sharing your suggestions and supporting the continued improvement of AI Assist.

Regards,
Janice G

Thanks, Janice. To help me plan internally, could you confirm:

  • Is this feature already in the product backlog with a rough priority level?

  • Is there a way for me to track its progress or vote/escalate through my account manager?

  • In the meantime, does tawk.to support any custom widget scripting or API access that would let me implement basic TTS myself?

If this isn’t on the near-term roadmap, I need to evaluate whether tawk.to’s AI Assist can meet my Q3 customer experience goals, so any clarity you can provide would be greatly appreciated.

I’ve made it much better. It works on mobile also, but your team should/could implement this within less than one work day. I could do this if you gave me access to your back-end.

Sincerely

Hi tawk.to Team,

I’m currently using a custom JavaScript workaround to add text-to-speech to my AI Assist bot messages. It works, but it’s not ideal. Here’s what I’m doing and what I need from you:

What I built:

  • A floating :loud_sound: button outside your widget

  • It captures bot messages via Tawk_API.onChatMessageAgent

  • Uses browser Web Speech API to read messages aloud

The problems with my workaround:

  1. The button sits outside the chat widget — not integrated with message bubbles

  2. Users have to click a separate button, not a speaker icon next to each message

  3. Positioning is hacky — I have to guess pixel offsets from your iframe

  4. Cross-origin iframe blocks me from injecting buttons directly into messages

What you need to do (minimal effort): Add a native speaker icon next to each AI Assist / bot message. When clicked, call:

window.speechSynthesis.speak(new SpeechSynthesisUtterance(messageText))

That’s it. The browser handles voices, speed, and playback. No API keys, no external services, no heavy lifting.

Why this matters:

  • Accessibility (screen readers aren’t enough)

  • Mobile users who don’t want to read long replies

  • Brand experience — voice makes the bot feel alive

  • Your competitors will add this eventually

My ask: Move TTS from “backlog” to “next sprint.” It’s a 20-line feature with massive UX impact.

Thanks, Richard Barton | Skan Digital
Founder & CEO
DigiDevice