Transcriptions and AI Insights

EnaCast automatically transcribes your show episodes and generates AI-powered insights such as titles, summaries, tags, and topic sections. This processing happens automatically after each recording.

Transcription Levels

Each show can be configured with a transcription level that determines how much audio is transcribed and what AI analysis is generated.

Level	Audio Transcribed	Analysis Generated
Full	Entire episode	Title, summary, tags, sections, highlighted quotes
15 minutes	First 15 min only	Title only
30 minutes	First 30 min only	Title only

When to Use Partial Transcription?

Partial transcription (15 or 30 minutes) uses less daily budget than full transcription (only the transcribed minutes are deducted, not the total episode duration). It’s useful when:

• You have long shows (sports broadcasts, council meetings) and only need an automatic title
• You want to take advantage of automatic start trimming without spending too much budget — 15 minutes is enough for the AI to detect where the actual program content begins
• You have a packed schedule and need to stretch your budget across more shows

What Each Level Generates

Full Transcription

With full transcription, EnaCast generates:

• Title: a descriptive title of the episode’s content
• Long summary: detailed description of the topics covered
• Short summary: condensed version for previews and social media
• Tags: relevant keywords to categorize the episode
• Topic sections: automatic markers that divide the episode by themes
• Highlighted quotes: noteworthy phrases extracted from the transcription
• Start and end detection: identification of where the actual content begins and ends, for automatic trimming

Partial Transcription (15 min / 30 min)

With partial transcription, EnaCast generates:

• Title: a title based on the first minutes of the episode
• No summaries, tags, or sections are generated (the content is incomplete)
• Episode end point is not detected (only the first minutes are available)

Configuring the Transcription Level

The transcription level is configured per show from the admin panel, in the show settings. When a new episode is recorded, it automatically inherits the level configured at that time.

Note

Changing a show’s transcription level only affects future episodes. Already transcribed episodes are not reprocessed.

Plans and Daily Limits

The transcription system has a daily limit of audio minutes that can be transcribed, depending on your plan.

Plan	Daily Limit	Best For
Basic	30 minutes/day	Stations with few daily shows
Pro	5 hours/day	Stations with a full schedule

The limit resets daily

Unused minutes do not carry over from one day to the next. The counter resets every day at midnight (UTC).

How Is the Budget Consumed?

The budget is consumed based on the duration of audio actually transcribed, taking the configured transcription level into account:

• A 60-minute episode with full transcription uses 60 minutes of the limit
• A 60-minute episode with 15-min transcription uses only 15 minutes of the limit
• A 60-minute episode with 30-min transcription uses only 30 minutes of the limit

The budget is reserved when the episode is marked for transcription, not when transcription completes.

Checking Your Budget

From the admin panel you can check:

• Today’s budget: how many minutes you’ve used and how many remain
• History: usage over the last 7-30 days
• Estimate: projected future usage based on shows with active transcription and their schedules

What Happens When the Budget Runs Out?

Important

When the daily limit is reached, episodes that finish recording that day are not transcribed and are not queued. They simply are not processed.

Episodes that are not transcribed on a given day are not automatically recovered the next day. There is no queue system that carries over pending episodes from previous days.

Priorities and Processing Order

Episodes are processed in the order they finish recording (or are manually trimmed). There is no priority system: the first one to finish is the first one to use the budget.

In practice, this means:

1. Shows that air earlier in the day are more likely to be transcribed, because budget is still available
2. Shows that air later in the day might not be transcribed if the budget has already been exhausted

Tip for stations with a full schedule

If you have more shows than your daily budget can cover, review which ones truly need automatic transcription. You can disable transcription for music shows or reruns to save the budget for the most relevant spoken content.

You can also use partial transcription (15 or 30 minutes) for shows where you only need an automatic title — this reduces both budget consumption and processing time.

Non-Transcribed Episodes

An episode that was not transcribed due to budget limits:

• Is still published on the website and RSS feed (with the usual wait times)
• Keeps the show’s title, description, and image as default values
• Does not have AI-generated summaries, tags, sections, or quotes
• Can be manually edited from the admin panel

Visual Summary

Automatic Start and End Trimming

EnaCast can automatically detect where the actual content of an episode begins and ends, removing silences, filler music, or pre-show content (such as news bulletins).

How It Works

When a show has the “Analyze start and end before cutting” option enabled in its settings:

1. The episode finishes recording
2. The AI analyzes the audio and detects the exact second where the program starts and where it ends
3. The episode is automatically trimmed using those points, with no editor intervention needed
4. The original audio is always preserved — only the trimmed version is replaced

Configuration

This feature is enabled per show from the admin panel. It requires the show to have transcription enabled, as the AI needs to analyze the audio to detect the trim points.

Automatic Trimming and Partial Transcription

Automatic trimming also works with partial transcription (15 or 30 minutes). This is an especially useful combination: the AI analyzes the first few minutes to detect the program start and trims it, using very little daily budget. End point detection is not available with partial transcription, since the analyzed audio doesn’t reach the end of the episode.

Result

Once automatically trimmed, the episode:

• Is published with clean audio (no silences or pre-show content)
• Appears in the RSS feed with reduced wait time (5 minutes instead of 30)
• Shows the AI-suggested trim points in the editor, in case the editor wants to adjust them manually

Frequently Asked Questions

Can I manually transcribe an episode that was not processed?

It is currently not possible to force transcription of a specific episode from the admin panel. If you need to transcribe an episode that was missed, contact technical support.

Can I change the transcription level of an already processed episode?

No. The transcription level is set at the time the episode is created. For future episodes, change the show’s configuration.

How long does transcription take?

The time depends on the episode duration and system load. Typically, a 60-minute episode is transcribed in a few minutes.

What languages does transcription support?

Transcription works in any language. The system automatically detects the audio language.

---

If you have questions, contact technical support at support@enacast.com