RTC Auto-Answer Bot

The calling system includes an RTC auto-answer bot for development and testing. This way you can iterate on the subsystem with a single phone. When you call the bot contact, the backend automatically notifies a remote bot server to join/leave the call.

Starting the Bot Server

It runs on Google Compute Engine (not Firebase) since it needs to be always running, listening for incoming calls.

1. SSH to the VM: gcloud compute ssh --zone "us-central1-c" "instance-20251031-20251031-055846" --project "pajama-dev"
2. Ensure no running server: ps aux | grep "videosdk_server.py"
3. To start the server for debugging: source venv/bin/activate && python videosdk_server.py
4. To start the server for long-term running:

source venv/bin/activate
nohup python videosdk_server.py > server.log 2>&1 &

Setting up a New VM

1. Reserve a VM from https://console.cloud.google.com/compute/instances. The bot uses ~0.15 vCPU per connection, so a micro should suffice. Ensure you "Add Firewall Rule" to allow IPv4 ranges: 0.0.0.0/0 and TCP Port: 58989.
2. SSH and copy over source files.
3. Install packages:

sudo apt update
sudo apt install software-properties-common -y
sudo apt install python3-pip python3.12-venv
python3 -m venv venv
source venv/bin/activate
pip3 install -r requirements.txt

4. Update the IP address in pajama/services/firebase/functions/pajama/rtc/operations/constants.py to the new VM's External IP.

Updating the video

1. Save your video in an optimized lower res format: ffmpeg -i loop.mp4 -vf "scale=180:320" -r 15 -c:v libx264 -preset ultrafast -crf 28 -b:v 150k loop_lowres.mp4
2. Copy file to VM
3. Restart server