Using Langchain with speech input

[jaymon0703]

Hi

Given Multimodal models are becoming more and more prevalent, is there any work on the Langchain roadmap for "reading" speech input natively, or should we use speech-to-text libraries for the foreseeable future?

Example is, i want to verbally ask a question of my app which utilizes a QA Chain, or API Chain or SQL Agent, etc. Being able to speak instead of type will add to productivity, and doing so natively is possible with direct access to the Gemini API for example.

Thanks!