How to Record Google Meet: Native Tools and Bot API Guide

You've just finished the most important customer call of the quarter. The recording? Didn't start. Maybe the host forgot to click the button. Maybe the account didn't have a Google Workspace license. Maybe the recording started but the transcript is buried in Google Drive with no programmatic way to retrieve it. Whatever the reason, you lost the data.

Recording meetings reliably, at scale, and in a way that feeds downstream systems is harder than it sounds. Native recording tools are built for individuals, not for applications that need to process hundreds of calls a week.

This guide covers both paths: using Google Meet's built-in recording features for straightforward cases, and using the MeetStream bot application programming interface (API) when you need programmatic control, automatic transcription, and webhook delivery.

In this guide, we'll cover how native Google Meet recording works and its limitations, how to deploy a recording bot via API, how to retrieve recordings and transcripts, and the webhook lifecycle from join to transcript. Let's get into it.

How Google Meet's native recording works

Google Meet's built-in recording feature saves meetings directly to Google Drive. When a recording ends, participants get an email with the Drive link and a copy is added to the meeting organizer's Drive folder.

The requirements are significant. You need Google Workspace (Business Standard, Business Plus, Enterprise, or Education Plus). Free Google accounts cannot record. The meeting host or co-host must manually start and stop the recording. Recordings capture the main shared view, not individual participant streams.

Accessing those recordings programmatically means navigating the Google Drive API. You'd query Drive for files with the meeting ID in metadata, download the MP4, then run it through a separate transcription pipeline. That's three different APIs, three sets of credentials, and a workflow that breaks the moment someone forgets to hit record.

For individual use, native recording is fine. For applications that need reliable, automatic recording across many meetings without human intervention, you need a different approach.

How recording bots work

A recording bot is a programmatic participant that joins a meeting, captures audio and video from inside the call, and makes the data available via webhooks and API endpoints. The bot appears as a named participant in the meeting, no browser automation or screen scraping involved.

The architecture is straightforward: you send an API request with the meeting URL, the bot joins within seconds, it captures the session, and when the meeting ends it processes and delivers the recording and transcript to your systems via webhook callbacks.

The advantage over native recording is control. You decide when a bot joins, what it captures, which transcription provider to use, and where the data goes. No host action required. No Google Workspace dependency. Works whether you own the meeting or not.

With MeetStream, Google Meet requires no additional platform setup. Unlike Zoom (which requires App Marketplace approval), you can start sending bots to Google Meet URLs immediately after getting your API key.

Recording Google Meet via the MeetStream API

The create bot endpoint is the entry point for everything. Send a POST request with the meeting URL and your configuration. The bot joins, records, and delivers data to your callback URL.

Here's a minimal example using curl:

curl -X POST https://api.meetstream.ai/api/v1/bots/create_bot \
  -H "Authorization: Token YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "meeting_link": "https://meet.google.com/abc-defg-hij",
    "bot_name": "MeetStream Recorder",
    "callback_url": "https://yourapp.com/webhooks/meetstream",
    "video_required": true,
    "recording_config": {
      "transcript": {
        "provider": {
          "name": "assemblyai"
        }
      },
      "retention": {
        "type": "timed",
        "hours": 24
      }
    }
  }'

The response gives you a bot_id and transcript_id you'll use to retrieve data later:

{
  "bot_id": "bot_abc123",
  "transcript_id": "txn_xyz789",
  "meeting_url": "https://meet.google.com/abc-defg-hij",
  "status": "Joining"
}

The recording_config block controls what gets processed after the meeting. Setting transcript.provider.name to assemblyai triggers AssemblyAI transcription once audio processing finishes. You can also use deepgram, jigsawstack, or meeting_captions. The retention.hours field sets how long recordings stay available for retrieval.

For scheduled meetings, pass a join_at parameter in ISO 8601 format. The bot queues and joins automatically three minutes before the specified time, no cron job needed on your end.

The webhook lifecycle for Google Meet recording

Once the bot is deployed, your callback_url receives a sequence of events. Understanding this flow is essential for building a reliable recording pipeline.

The sequence is: bot.joining fires when the bot attempts to enter the lobby. bot.inmeeting fires when the bot is admitted and recording begins. bot.stopped fires when the meeting ends or the bot is removed. Then, asynchronously, audio.processed, video.processed, and transcription.processed fire as each artifact becomes available.

// bot.stopped payload example
{
  "event": "bot.stopped",
  "bot_id": "bot_abc123",
  "transcript_id": "txn_xyz789",
  "bot_status": "Stopped",
  "timestamp": "2026-04-02T14:30:00Z"
}

// transcription.processed payload example
{
  "event": "transcription.processed",
  "bot_id": "bot_abc123",
  "transcript_id": "txn_xyz789",
  "status": "completed",
  "timestamp": "2026-04-02T14:35:00Z"
}

The bot_status field in bot.stopped tells you why the bot left. Stopped means a clean exit. NotAllowed means the host didn't admit the bot from the waiting room. Denied means the bot was kicked. Error means something went wrong on the infrastructure side. Build your error handling around these states.

Wait for transcription.processed before trying to retrieve the transcript. Processing time depends on meeting length and the transcription provider you chose.

Retrieving the recording and transcript

Once you receive the processed webhooks, retrieve artifacts using the IDs from the initial create_bot response.

For the transcript:

curl -X GET https://api.meetstream.ai/api/v1/transcript/txn_xyz789/get_transcript \
  -H "Authorization: Token YOUR_API_KEY"

The response returns speaker-diarized text with timestamps, ready to feed into summaries, CRM updates, search indexes, or any downstream system your application needs.

Video and audio are retrieved via the bot ID using separate endpoints (GET /bots/{bot_id}/video and GET /bots/{bot_id}/audio). Set video_required: false if you only need audio and transcription, which reduces processing time and storage.

Scheduling bots with Google Calendar integration

For teams that want fully automatic recording without any manual API calls, MeetStream's Google Calendar integration handles scheduling automatically. Connect your calendar, and MeetStream detects upcoming meetings, joins three minutes before the start time, and delivers recordings and transcripts without any per-meeting API calls.

Real-time change detection means if a meeting is rescheduled or cancelled, the bot adjusts accordingly. This is the path most teams take once they've validated the basic recording flow.

Limitations and what to watch for

A few practical constraints to keep in mind. The bot appears as a participant named whatever you set in bot_name. Meeting participants can see it in the attendee list. For most business use cases this is fine and expected, but some meeting hosts may remove it if they're not expecting an automated participant. Setting a recognizable bot name reduces confusion.

If the meeting host requires explicit admission from the waiting room and nobody admits the bot, you'll get a NotAllowed status. Build retry logic or alerting for this case if your use case involves meetings you don't control.

Google Meet's end-to-end encryption settings can affect bot behavior. Standard workspace meetings work fine. Meetings with enhanced encryption enabled may block third-party participants.

Transcription processing is asynchronous and typically completes within a few minutes of the meeting ending. For very long meetings (3+ hours), allow more time before polling or acting on transcription.processed events.

How MeetStream fits in

MeetStream provides the recording bot infrastructure so you don't have to build or maintain it. Send a single API request with a Google Meet URL, get back a transcript and recording via webhook. No additional platform credentials needed for Google Meet, no screen scraping, no headless browser maintenance. It works the same way across Zoom and Microsoft Teams, so if your application needs to handle meetings across platforms, the same API call handles all three.

Conclusion

Native Google Meet recording covers simple, human-driven workflows. For applications that need to record Google Meet sessions programmatically, without depending on host actions or Google Workspace licenses, the bot API approach is more reliable and more scalable. The key pieces are: a POST to create a bot with a callback_url and recording_config, webhook handling for the bot lifecycle events, and GET requests for transcript and media once processing completes. Google Meet requires no additional setup with MeetStream, making it the fastest platform to get started with.

Get started free at meetstream.ai or see the full API reference at docs.meetstream.ai.

Frequently Asked Questions

How do I record Google Meet without a Google Workspace account?

Google Meet's native recording requires a Google Workspace subscription. Without it, the record button simply isn't available. Using a recording bot API like MeetStream bypasses this requirement entirely. The bot joins the meeting as a participant and captures audio and video without needing any special Google account permissions.

Does the Google Meet recording bot API require any special setup?

MeetStream's Google Meet integration requires no platform-specific setup. Unlike Zoom, which requires App Marketplace approval, you can send bots to Google Meet URLs immediately after getting an API key. Just pass the meeting URL in your create_bot request and the bot handles the rest.

How long does it take to get a Google Meet transcript after the meeting?

Transcription processing typically completes within two to five minutes of the meeting ending for standard-length calls. You'll receive a transcription.processed webhook event when it's ready. For meetings longer than two hours, allow up to ten minutes for processing to complete before retrieving the transcript.

Can I record a Google Meet that I'm not hosting?

Yes, the recording bot joins as a participant, so it can record meetings regardless of whether you're the host. The bot appears in the participant list with the name you set in bot_name. If the meeting uses a waiting room, someone with host permissions will need to admit the bot, or you can disable waiting room requirements on your end if you control the meeting settings.

What transcription providers are available for Google Meet recording?

MeetStream supports AssemblyAI, Deepgram, JigsawStack, and meeting_captions for post-call transcription. For real-time transcription during the meeting, Deepgram Streaming and AssemblyAI Streaming are available via the live_transcription_required parameter. Each provider has different accuracy and latency characteristics, so the best choice depends on your specific use case.