Interactive Podcast Audio Integration Tool

Objective

Build a standalone Windows script that:

Plays a pre-recorded audio file into NotebookML’s mic input (using VB-Cable)

Records the browser output (Chromium) for 60 seconds and saves it as a .wav file

Tech Requirements

OS: Windows 10 or 11

Language: Python (preferred) or JavaScript (Node.js)

Tools:

VB-Cable (A+B) for audio routing

FFmpeg for recording audio

(Optional) Playwright to simulate “Join” click in NotebookML

Functionality

Step 1 – Audio Injection

Play a specified .wav file (e.g., question.wav) into the Chromium browser’s microphone input via VB-Cable

The browser should already be open at https://notebooklm.google.com/ in Interactive Mode

Ideally, wait 5 seconds before playback to allow time for the host's welcome message and for the podcast to enter listening mode

Step 2 – Audio Recording

Start recording output from Chromium’s speaker output via VB-Cable

Save a .wav file (e.g., response_2025-05-30T12-00-00.wav)

Duration: 60 seconds (should be configurable in the script)

Step 3 – Output

Save the .wav file to the ./recordings/ folder

Print status to console only:

Injecting audio...

Recording answer...

Saved response to recordings/response_.wav

Configuration (can be hardcoded)

Input audio file path

Output WAV file location

Recording duration (default: 60s)

Device names for VB-Cable input/output (optional to expose)

Deliverables

A working Python or JavaScript script file (inject_and_record.py or .js)

README with setup instructions and list of dependencies

All source code in editable form, tested on Windows 10 or 11 using VB-Cable and Chromium