VideoDB Documentation

Pages

Meeting Recording Agent Quickstart

What you’ll build

A headless “bot-participant” that joins your Zoom, Meet, or Teams call, records high-quality audio + video, and upload straight into your VideoDB collection.

Why it matters

Once the recording lands in VideoDB, you immediately unlock:

Semantic search & clipping – jump to moments in seconds and watch instantly.

LLM summaries – generate follow-up emails or full minutes with one prompt.

Screen-share parsing – scrape on-screen data for deeper analytics.

Workflow hooks – auto fill CRM fields, trigger Zapier actions, connect n8n nodes or raise Jira tickets.

AI Interfaces - chat with

Director⁠

agent to record and

get tailored insight from your meetings⁠

MCP Server - use

VideoDB MCP Server⁠

to control your meetings from anywhere.

Build advanced workflows :

CRM auto update – create & enrich deals, push next steps.

Candidate interview scoring – extract skills, sentiment, red flags.

Webinar chaptering & highlights – publish shareable clips in minutes.

(Feel free to fork—each notebook uses the same SDK primitives you’ll wire into production.)

⁠

⁠

Platforms : Google Meet, Teams, Zoom.

Recorder and Post Meeting Analysis

Key Features:

⁠

Brand-able Participant - Configure a custom bot name and avatar for a seamless, white-label presence in every call.

Live Status & Completion Events - Fetch recording state at any time, and receive a final “recording-complete” callback with file URLs and metadata.

Speaker-Timeline Metadata - Per-speaker start/stop timestamps are returned for Google Meet only. Microsoft Teams and Zoom support ship next. Use this timeline to align transcripts, analytics, or highlights from individual speakers.

Getting Started

⁠

Install the SDK

!pip install --upgrade videodb

import os

os.environ["VIDEO_DB_API_KEY"] = <Your API Key>

Record a Meeting

⁠

import videodb

conn = videodb.connect(api_key=os.getenv("VIDEO_DB_API_KEY"))

meeting = conn.record_meeting(

meeting_url="https://meet.google.com/abc-defg-hij",

bot_name="Your Bot Name", # Name displayed in the meeting

bot_image_url = "https://link-to-image.jpg" # Image displayed in the meeting

meeting_title="Your Meeting Title", # Identifies the meeting internally

callback_url="https://yourapp.com/webhook",

callback_data={"internal_id": "123"}, # Your custom metadata

time_zone="America/Los_Angeles"

)

print(f"Meeting recording initiated with ID: {meeting.id}")

You can also use collection object to record and upload meeting on that collection.

my_coll = conn.get_collection("your-collection-id")

meeting = my_coll.record_meeting(

meeting_url="https://zoom.us/j/123456789",

bot_name="Team Recorder Bot",

meeting_title="Weekly Standup",

callback_url="https://yourapp.com/webhook"

)

Tracking Recording Status

⁠

meeting.refresh()

print("Current status:", meeting.status)

Wait for completion (recommended method):

# Wait for up to 1 hour, polling every 60 seconds

if meeting.wait_for_status("done", timeout=3600, interval=60):

print("Recording successful.")

else:

print("Recording encountered an issue.")

Recording status:

initializing - Bot is being set up

processing - Bot is actively recording the meeting

done - Recording completed successfully

failed - Recording failed

Callback Notifications

⁠

VideoDB sends a webhook (HTTP POST) to your specified callback_url when recording completes. You can use these to automate post-processing tasks.

Success:

{

"success": true,

"message": "Meeting recording completed.",

"data": {

"video_id": "vid_123",

"speaker_timeline": [{"speaker_name": "Alice", "start_time_seconds": 9.94}],

"stream_url": "...",

"player_url": "..."

}

Failure:

{

"success": false,

"message": "Failed to record meeting."

}

Accessing the Recorded Video

⁠

Retrieve your video directly from connection conn.get_meeting(”xx-xx”) or collection object after completion:

meeting = coll.get_meeting("xx-xx-xx")

print(meeting.id)

print(meeting.video_id)

# Get video object

video = collection.get(meeing.video_id)

Meeting Object

Meeting Object

ID of the meeting to get it from conn and collection objects

bot_name

Name of the bot that appears in the meeting.

bot_image_url

Image of the bot that appears in the meeting.

meeting_url

URL of the meeting being recorded

meeting_title

Title of the meeting

collection_id

ID of the collection where the recording will be stored

status

Current status of the recording (see Status Values below)

time_zone

Time zone of the meeting

is_active

Property that returns True if meeting is currently being recorded

is_completed

Property that returns True if meeting recording is done

speaker_timeline

Speaker timeline.

video_id

Video ID of the meeting.

There are no rows in this table

⁠

Stay tuned for upcoming features 🙌🏼

⁠

⁠

Meeting Scheduling: Schedule recordings in advance with the join_at parameter or calendar connects.

Real-Time Stream Access: Access audio and video streams live during meetings.

Live Action Integration: Connect real-time meeting streams directly into

Real‑Time Video Pipeline⁠

Audio Responses: Have your bot respond back to participants using conversational intelligence.

Human-like Screen Sharing: Automate dynamic, interactive screen-sharing use cases.

⁠

You're all set to start automating your meeting recordings with VideoDB! If you encounter any issues or have questions, reach out via our support channels on

Discord

⁠