I’d like to suggest a new AI-powered feature for ONLYOFFICE Docs and PDF modules: Please consider integrating text-to-speech (TTS) capabilities, allowing users to have documents (DOCX and PDF, either in full or partially) read aloud via services like ElevenLabs – known for its high-quality, natural-sounding voices. Ideally, users should be able to connect their ElevenLabs account via API key, just as they can with your current AI plugins, and also have the option to connect other TTS providers like OpenAI or Google if desired.
The voice quality and usability should match the ElevenLabs Reader app experience available on Android and iOS – clear, expressive, and easy to control. This would make ONLYOFFICE much more accessible and efficient for users who rely on listening to documents, for productivity or accessibility reasons.
Thanks for considering this – it would be a real game-changer!
Hello @bavariar
Thank you for the suggestion. Considering that AI-powered features become more and more popular, I’d like to ask you to elaborate a little on your suggestion – in which form you would like to see TTS capabilities being implemented?
Currently, we have AI plugin that allows connecting various AI providers (even local ones) to assist you with document editing, there is also simple Speech plugin that allows reading selected text aloud. The first one does not bring TTS capabilities as of now, but the second one is quite simple and has no AI features in it.
Hi Constantine,
How to implement it, I leave it to the programming team. But the expectation is that your very simple speech plug-in should be enhanced by the services of Eleven Labs. The quality of the speech of the speech plug-in for complicated documents is very limited. The requirement would be to provide some kind of a connection between OnlyOffice and Eleven Labs to be able to use the very advanced text-to-speech algorithms of Eleven Labs in order to be able to read aloud any Office document, be it a PDF or a docx document.
How to do the implementation is a question of technology feasibility, but probably either it could work via an MCP client-server architecture or maybe an API key toward Eleven Labs.
Thank you for the reply. I believe that TTS capabilities, mostly AI-powered ones, belong to the AI plugin. That said, we will discuss possibility to include TTS as one of the options, alongside chat, translation, etc., of the AI plugin in future versions.
Please await for my feedback.
UPD: enhancement to add TTS option to the list of AI plugin feature has been registered. Once it becomes available, I will inform you.
THank you for your suggestion.