node-quickstart

Node.js Capture Quickstart

A complete example showing real-time media capture, indexing, and transcription using the VideoDB Node.js SDK.

Overview

This app runs as a single script that demonstrates the full VideoDB Capture workflow:

• Creates a capture session with WebSocket for real-time events
• Captures screen, mic, and system audio using the native capture client
• Performs live transcription and audio/visual indexing on all streams
• Prints real-time results to the console

Prerequisites

1. Node.js 18+
2. VideoDB API Key: Get one from console.videodb.io

Setup

1. Install Dependencies:

   npm install

2. Configure Environment: Copy .env.example to .env and add your API key:

   cp .env.example .env
   # Edit .env and set VIDEODB_API_KEY

Running the App

npm start

The app will:

1. Connect to VideoDB and create a capture session
2. Request permissions for microphone and screen capture
3. Start recording mic, system audio, and screen
4. Start transcription and indexing on all streams
5. Print real-time results via WebSocket

Press Ctrl+C to stop.

What's Happening?

• The script creates a capture session and generates a client token
• The native capture client discovers channels (mic, display, system audio)
• Channels with record: true save the recording to VideoDB after capture stops
• Once active, transcription and audio indexing start on both audio streams (mic + system audio)
• Visual indexing starts on the display stream
• Real-time results stream back via WebSocket

Expected Output

============================================================
VideoDB Capture - Node.js Quickstart
============================================================

Connecting to VideoDB...
Using collection: c-xxx
Connecting WebSocket...
WebSocket connected: ws_id
Creating capture session...
Session created: cap-xxx
Client token generated

Requesting permissions...
Discovering channels...
  - mic:default (audio): Default Microphone
  - display:1 (video): Built-in Display
  - system_audio:default (audio): System Audio

Starting capture with 3 channel(s):
  - mic:default
  - display:1
  - system_audio:default
Capture started!

Waiting for session to become active...
Session status: active
  Starting audio indexing on: rts-xxx
  Audio indexing started: rts-xxx
  Starting audio indexing on: rts-yyy
  Audio indexing started: rts-yyy
  Starting visual indexing on: rts-zzz
  Visual indexing started: rts-zzz

============================================================
Recording... Press Ctrl+C to stop
============================================================

[Transcript:mic] real-time transcription streams live...
[Transcript:system_audio] real-time transcription streams live...

**************************************************
[Audio Index:mic] Summary of what is being discussed
**************************************************

**************************************************
[Audio Index:system_audio] Summary of what is being discussed
**************************************************

**************************************************
[Visual Index:screen] One-sentence description of screen content
**************************************************

Stopping the Recording

Press Enter to stop recording gracefully, or Ctrl+C to force quit. The app will:

1. Stop the capture session
2. Shut down the native binary
3. Close the WebSocket connection

Wait a few seconds for the session to finalize. Once exported, you can view the video in the VideoDB console.

Environment Variables

Variable	Required	Description
VIDEODB_API_KEY	Yes	Your VideoDB API key from console.videodb.io
VIDEODB_COLLECTION_ID	No	Collection ID (defaults to default)

Troubleshooting

API Key Issues

• Error: VIDEODB_API_KEY is required
  • Make sure you've created a .env file with your API key
  • Verify the API key is valid at console.videodb.io

Permission Errors

• Error: Permission request failed
  • Grant system permissions for microphone and screen recording when prompted
  • On macOS: Check System Settings > Privacy & Security

No Channels Available

• The native capture binary may not be running properly
• Try reinstalling: npm install

No AI Results Appearing

• Make sure you're generating audio or visual activity to be indexed
• Verify the capture session status shows as "active"
• Check that the WebSocket connection was established successfully