Handling Notetaker media files

After your Notetaker bot leaves a meeting, Nylas automatically creates three types of media for the event: an MP4 audio/video recording, a PNG thumbnail from the video, and a text transcript of the conversation. Nylas also sends you notetaker.media webhook notifications with updates about the media processing status.

When Nylas is finished processing the files, it sends a notetaker.media webhook notification with URLs for the recording, thumbnail, and transcript.

Nylas stores Notetaker media files for a maximum of 14 days. After the 14-day storage period, the files are deleted and you can’t retrieve them using the Nylas API. We recommend downloading and storing these files as soon as they’re available if you need access to them beyond Nylas’ retention period.

Download and store Notetaker media files

Nylas only stored Notetaker media files for up to 14 days, so we recommend you set up a system that downloads and stores the files as they become available. The easiest way to do this is to subscribe to notetaker.media webhook notifications and set up an automated process to download the files when their state is available.

The media URLs included in notetaker.media webhook notifications are valid for 60 minutes. After that point, you’ll need to make a Download Notetaker Media request to retrieve the files.

{
  "specversion": "1.0",
  "type": "notetaker.media",
  "source": "/nylas/notetaker",
  "id": "<WEBHOOK_ID>",
  "time": 1737500935555,
  "webhook_delivery_attempt": 1,
  "data": {
    "application_id": "<NYLAS_APPLICATION_ID>",
    "object": {
      "object": "notetaker",
      "grant_id": "<NYLAS_GRANT_ID>",
      "id": "<NOTETAKER_ID>",
      "state": "available",
      "media": {
        "recording": "<RECORDING_URL>",
        "thumbnail": "<THUMBNAIL_URL>",
        "transcript": "<TRANSCRIPT_URL>"
      }
    }
  }
}

You can then store the files in your own infrastructure and present them to your users as necessary.

Notetaker media file formats

Nylas automatically creates three types of media for each event your Notetaker records: MP4 audio/video recordings, PNG thumbnails, and text transcripts.

Notetaker video recording format

Notetaker video recordings capture the main meeting interface, excluding any browser UI elements and meeting platform toolbars. The behavior varies slightly for each provider:

• Google Meet: Recording of participant grid and shared screen content.
• Microsoft Teams: Recording of the main meeting area.
• Zoom: Recording of the full meeting window.

Nylas generates an MP4 file with the following specifications:

• Video specs:
  • Format: .mp4 video file
  • Resolution: 1280x720
  • Frame rate: 12 FPS
• Audio specs:
  • Format: Stereo audio track.
  • Quality: 48 kHz sampling.
  • Content: All meeting audio, including participants and shared content.

Notetaker thumbnail format

Nylas automatically generates a thumbnail from approximately half-way through the meeting’s video recording with the following specifications:

• Format: .png image file
• Resolution: 1280px width (height scaled proportionally)

Notetaker transcript format

For most events, Nylas includes the following information in each transcript:

• Speaker attribution: "Speaker 1", "Speaker 2", and so on.
• Timing information: start and end times, in milliseconds.

It also separates each speaker’s contribution into individual text segments that align fall within the start and end times, as in the following example.

{
  "object": "transcript",
  "type": "speaker_labelled",
  "transcript": [
    {
      "speaker": "Speaker 1",
      "start": 100,
      "end": 10420,
      "text": "Did you know that a day on Venus is longer than its year? It takes Venus about 243 Earth days to rotate once, but only about 225 Earth days to orbit the Sun."
    },
    {
      "speaker": "Speaker 2",
      "start": 10500,
      "end": 12500,
      "text": "That's wild. So technically, you could have a birthday before a sunrise there."
    }
  ]
}

In rare cases, Nylas might return the transcript as a raw text file without speaker labels or timestamps.

{
  "object": "transcript",
  "type": "raw",
  "transcript": "The Moon is slowly moving away from the Earth at a rate of about 3.8 centimeters per year."
}

Although it’s rare that Nylas returns transcripts as raw text, your project should be able to handle both formats. You can use the type field to determine which format you’re working with.

Keep in mind

Keep the following information in mind as you work with Notetaker media files:

• Nylas takes a few minutes to process media files after Notetaker leaves a meeting.
• We recommend you add error handling and retry logic to your project for downloading media files, in case the initial download fails.
• Make sure your project can handle multiple downloads simultaneously, and plan for scale; popular applications might need to handle thousands of downloads per day.
• If you’re storing files to present them to your users, be sure to…
  • Store the files securely and create retention policies that meet your business needs.
  • Back up the stored files regularly.