Review meetings from external platforms

If your call recording tool isn’t natively supported by Triple Session, you can still analyze your calls using an incoming webhook. Send a transcript or recording URL to AI Coach and it will automatically review the meeting using the appropriate scorecard, whether it’s a sales call transcript, a WhatsApp conversation, or a recording from an internal system. This approach uses Triple Session’s Automations feature with a Webhook Received trigger and a Review Meeting action, coordinated through a tool like Zapier.

Supported file types

AI Coach accepts the following formats when uploading transcripts manually or via webhook:

.txt

Plain text transcript files

.vtt

WebVTT transcript files (recommended for accuracy)

.srt

SRT subtitle/transcript files

.mp4

Video recording files

.mp3

Audio recording files

.wav

Audio recording files

.m4a

Audio recording files

Set up webhook-based reviews

Connect your webhook in Triple Session

In Triple Session, go to Settings → Integrations → Connect to access the Automations webhook setup.

Webhook setup in Settings > Integrations

Configure the automation

Create an automation using the Webhook Received trigger connected to the Review Meeting action.

Automation with Webhook Received trigger and Review Meeting action

Copy the webhook credentials

Open the trigger. Copy the Webhook URL, the Webhook secret, and the values in Headers to send. With Signed or API key selected, send the secret in the x-ts-automation-api-key header. Paste these into your automation tool (such as Zapier) to authenticate requests.

Webhook trigger settings with the Endpoint, the webhook secret and the headers to send

Once the webhook receives a payload, AI Coach automatically reviews the meeting using the appropriate scorecard.

Webhook payload schema

The automation webhook expects a JSON payload (content-type: application/json) with the following structure. Form-encoded bodies (application/x-www-form-urlencoded) are also accepted for the plain text and number fields. custom_metadata needs a JSON body. The trigger’s Request body section lists the same fields, with a JSON tab and an example request you can copy.

Request body fields in the webhook trigger settings

Required fields

sales_rep_email (string) — the email address of the sales representative. Must match a Triple Session user email in your company.

Meeting content

Provide at least one of these fields:

transcript_body (string) — the full transcript content of the meeting.
recording_url (string) — the http(s) URL of the recording in mp4, mp3, m4a, or wav format. It can be a public link or a file on a Recording source protected by a credential.

Optional fields

meeting_id (string) — unique identifier for the meeting. Strongly recommended to prevent duplicate reviews.
meeting_name (string) — name or title of the meeting. If omitted, AI Coach generates a name automatically.
meeting_date (string) — original meeting date.
meeting_duration_in_minutes (number) — duration of the meeting in minutes.
custom_metadata (object) — optional JSON object with context from your external system. Use it to send routing hints such as call type, CRM object type, direction, stage, pipeline, or source platform.

Meeting date examples

Send a date-only value when you only know the meeting’s calendar date:

{
  "meeting_date": "2026-06-20"
}

Date-only values are stored at noon UTC, so the evaluation date stays on the intended calendar day. Send an ISO datetime with a timezone offset when you know the exact local meeting time:

{
  "meeting_date": "2026-06-20T15:30:00-03:00"
}

Send an ISO datetime with Z when you already have the exact UTC time:

{
  "meeting_date": "2026-06-20T18:30:00Z"
}

Do not send a datetime without a timezone, such as 2026-06-20T15:30:00.

Custom metadata

Use custom_metadata when the source platform knows context that is not obvious from the transcript or recording. For example, you can send whether the call was a demo, renewal, discovery call, or outbound conversation. When the Review Meeting action uses smart scorecard selection, AI Coach uses this metadata as extra context to choose the best matching scorecard. If the action uses a fixed scorecard, the metadata is accepted but does not change scorecard selection. The value must be a JSON object. Keep it concise. Triple Session accepts up to 10,000 characters after the metadata object is serialized.

Example payload

{
  "sales_rep_email": "john.doe@company.com",
  "meeting_id": "meeting_12345",
  "meeting_name": "Q4 Sales Review with Acme Corp",
  "meeting_date": "2026-06-20",
  "meeting_duration_in_minutes": 45,
  "transcript_body": "John: Good morning, thank you for joining us today...",
  "recording_url": "https://example.com/recording-123",
  "custom_metadata": {
    "call_type": "demo",
    "crm_object_type": "deal",
    "direction": "outbound",
    "stage": "evaluation"
  }
}

Authenticate requests

Every request sends two headers with values from the trigger’s Headers to send:

Header	Value
`x-ts-automation-id`	The automation ID.
`x-ts-automation-api-key`	The webhook secret.

This works with any tool that can set fixed headers, such as Zapier. If your own code sends the requests and your security policy doesn’t allow a shared secret in request headers, you can sign each request instead. See Signed requests.

Recordings behind authentication

By default, Recording source authentication on the Review Meeting action is set to None, and Triple Session downloads recording_url without credentials. A public or pre-signed link works as is. If your recordings aren’t public, Triple Session can download each file with a Basic, Bearer or API key credential that your admin enters in the action. See Recordings behind authentication for setup, the Recording source contract and how the credential is protected.

Response codes

Code	Meaning
`200`	Accepted. The review runs in the background. The response says `"Meeting already processed"` when the same `meeting_id` was accepted in the last 24 hours.
`400`	Invalid body or payload data. The `error` field says what’s wrong.
`401`	Missing or wrong API key or signature, timestamp more than 5 minutes off, signed requests required, or the automation is paused or isn’t a Webhook Received automation. The `error` field says which.
`413`	Body larger than 1 MB.
`429`	Rate limit reached (100 requests per hour per automation by default). Retry after the number of seconds in `Retry-After`.
`500`	Temporary error. Retry with backoff.

Additional notes

The sales_rep_email must match a Triple Session user email that belongs to your company.
Sending a value in meeting_id is strongly recommended — it prevents duplicate reviews and adds idempotency to the endpoint. The same meeting_id is accepted once per 24 hours.
If a request returns 200 but no run appears in the automation’s Logs, the sales_rep_email doesn’t match a member of your company.
The transcript body must be a string in the transcript_body field. VTT format is recommended for the best accuracy.
The recording URL must point to a publicly downloadable file, or to a Recording source protected by a credential you enter in the automation (details). Supported formats are mp3, mp4, m4a, and wav, up to 2 GB.
custom_metadata should contain short, structured values that help classify the meeting. Don’t send full transcripts, recordings, or large notes in this field.

Download transcripts from supported platforms

If you need to export transcripts from Zoom or Fireflies to upload manually, follow these steps.

Download a transcript from Zoom

Go to Recordings in your Zoom account

Visit the recordings section of your Zoom account to find all available cloud recordings.

Download the recording files

Find the recording you want to analyze and click Download. This downloads all associated files, including the transcript.

Locate the .vtt file

Your download will include a file with a .vtt extension. This is the transcript file AI Coach uses for analysis.

Upload to AI Coach

Go to AI Coach, upload the .vtt transcript file, fill out the review details (rep name and scorecard), and click Start Review.

Download a transcript from Fireflies

Open Meetings in Fireflies

In your Fireflies account, click Meetings.

Select a meeting

Choose the meeting you want to export.

Download the transcript

From your Notepad, click Download, then select Download Transcript.

Choose SRT format

Select the SRT format. Enable the timestamp and speaker toggles before downloading.

Upload to AI Coach

In AI Coach, start a new review. Name the rep, select a scorecard, upload the SRT file, and click Start Analysis.

Get Started

AI Coach

AI Coach Bot

Integrations

CRM Sync

Workspace

Automations

Academy

Review meetings from external platforms

Supported file types

.txt

.vtt

.srt

.mp4

.mp3

.wav

.m4a

Set up webhook-based reviews

Webhook payload schema

Required fields

Meeting content

Optional fields

Meeting date examples

Custom metadata

Example payload

Authenticate requests

Recordings behind authentication

Response codes

Additional notes

Download transcripts from supported platforms

Download a transcript from Zoom

Download a transcript from Fireflies

​Supported file types

.txt

.vtt

.srt

.mp4

.mp3

.wav

.m4a

​Set up webhook-based reviews

​Webhook payload schema

​Required fields

​Meeting content

​Optional fields

​Meeting date examples

​Custom metadata

​Example payload

​Authenticate requests

​Recordings behind authentication

​Response codes

​Additional notes

​Download transcripts from supported platforms

​Download a transcript from Zoom

​Download a transcript from Fireflies

Supported file types

Set up webhook-based reviews

Webhook payload schema

Required fields

Meeting content

Optional fields

Meeting date examples

Custom metadata

Example payload

Authenticate requests

Recordings behind authentication

Response codes

Additional notes

Download transcripts from supported platforms

Download a transcript from Zoom

Download a transcript from Fireflies