Feature Request: Support for programmatically injecting/intercepting WebRTC audio streams in call.offer and call.accept

[uriva]

Describe the feature you'd like
Currently, the WPP.call module supports basic call controls such as WPP.call.offer, WPP.call.accept, and WPP.call.reject. While this allows triggering the browser's native call interface (dialing/answering), there is currently no built-in way to programmatically read or write the raw media streams (such as audio/video).

It would be highly valuable to have a way to programmatically:

1. Inject a custom audio stream (e.g., from a virtual microphone or raw buffer) into an outgoing call initiated by call.offer or an accepted call via call.accept.
2. Intercept and extract the incoming remote audio stream (the caller's voice) as a raw audio buffer/stream.

Under the hood, this could theoretically be achieved by:

• Overriding/hooking into navigator.mediaDevices.getUserMedia when a call is made to inject a custom MediaStream (e.g., controlled via Web Audio API).
• Intercepting the remote MediaStreamTrack from the active RTCPeerConnection instance and piping the raw audio chunks (e.g., PCM) back.

Would you accept a PR?
I am interested in exploring a potential solution for this. Before diving into implementation details or submitting code, I wanted to check:

1. Is this a feature you would be open to accepting a PR for?
2. Have you already explored or discussed a specific architectural approach for bridging live media streams between the Node/Puppeteer layer and the injected client-side script?

[dev-robin-hood]

I’m not sure this belongs in wa-js.

WPP.call already handles the call control layer (offer, accept, reject). Audio injection and remote media stream interception seem to be a separate concern that could be handled entirely outside the library, at the browser/WebRTC level.

Wouldn’t it make more sense to keep this functionality outside of wa-js rather than adding media-stream handling to the library itself?

[uriva]

if possible - yes
you mean the API will let me trasnmit audio input/output from a remote server as is?

[dev-robin-hood]

No... I'm saying you can do this without needing to implement it in the library.

As I understand it, you want wa-js to handle WhatsApp calls from external APIs—not necessarily the official WhatsApp service, but rather something like Baileys, for example?

[dev-robin-hood]

What do you think about this? @manfe

[uriva]

@dev-robin-hood i'm not sure what you mean still
I don't really have an easy way to send commands to the browser wa-js is using
iiuc wa-js reverses the whatsapp app so sending/receiving audio sounds well within its responsibility