Skip to content
Notifications
Clear all

Am I the only one who finds the highlight feature completely useless?

1 Posts
1 Users
0 Reactions
0 Views
(@integration_ian_2)
Reputable Member
Joined: 2 months ago
Posts: 203
Topic starter   [#22348]

Okay, I have to get this off my chest because I've been wrestling with it for months across dozens of meeting transcriptions. I've tried to incorporate Otter.ai's highlight feature into my automated workflows, and I've concluded it's borderline unusable for any serious integration or data extraction purpose. It feels like a half-implemented gimmick rather than a powerful tool.

My core issue is the complete lack of structure and actionable data. When you create a highlight, what are you actually creating? It's just a visual tag in the transcript with no corresponding metadata that can be reliably accessed. Let me break down the specific problems I've hit when trying to build automations around it:

* **No API Support for Highlights:** This is the biggest blocker. The Otter.ai API, which I've used to pull transcripts into my CRM and project management tools, has **zero endpoints for retrieving highlights**. You can fetch the transcript text, but the highlights are simply not present in the JSON response. This means any automation that depends on finding "the important parts" flagged by a user is dead on arrival.
* **No Export Functionality:** Even manually, you can't export a clean list of just your highlighted sentences. The "Summary" that includes highlights is a proprietary view within Otter. If you want to send just the key points to a Notion doc or a Slack channel, you're forced to copy-paste manually.
* **Highlights Don't Create Discrete Data Objects:** In a well-designed system, highlighting a sentence would create a timestamped, text-snippet object with a unique ID. You could then trigger a webhook to send that snippet elsewhere. In Otter, it's just a yellow mark on a static page.

For example, here's what I *wanted* to build in Make.com (formerly Integromat):

```json
// Ideal Scenario: API returns highlights array
{
"transcript_id": "abc123",
"highlights": [
{
"id": "h1",
"text": "We need to prioritize the Q3 integration launch.",
"timestamp": "00:12:45",
"created_by": "user@email.com"
}
]
}
```

With that structure, I could easily filter and route each highlight to a different tool. But since the data isn't there, the entire feature is siloed within Otter's UI.

Am I missing something? Has anyone found a workaround to actually *use* highlights in an automated, API-first manner? Or do we all just agree it's a superficial feature that looks good in a demo but provides zero utility for building connected systems?

api first


api first


   
Quote