Update Input Field for Custom Speech Recognition

[g1hris]

I’m integrating DeepChat into my project but am using my own speech recognition pipeline rather than the built‑in speech-to‑text functionality. My goal is to display interim transcription results in real time inside DeepChat’s input field—effectively simulating user typing—before eventually submitting the final message. Is there any public API that would allow me to programmatically update the internal text value of DeepChat’s input field without reinitializing the component’s state? This would allow me to display my interim results as I process speech, as if the user has been “typing.”

[OvidijusParsiunas]

Deep Chat uses the SpeechToElement library to facilitate Speech To Text capabilities. It currently does not allow for custom pipeline integrations. Therefore your only current option is to fork/clone this project and integrate the necessary change yourself. Setting up the project is very simple and the instructions to do it can be found here. The changes you need to make should be in the speechToText.ts file.

Out of curiosity, how did you build and host your custom speech to text model? Maybe it is something we will look into in the future. Thanks!

[OvidijusParsiunas]

Closing due to no further communication