---
title: "Session summary: one button, paid tier | VoxisLive"
description: "A recap of a finished session, written when you press the button. Paid tier, never automatic, in the transcript's own language."
url: https://voxislive.com/features/session-summary
language: "en"
source_html: https://voxislive.com/features/session-summary
site: "Voxis — real-time voice translation"
og_image: https://voxislive.com/assets/og.png
generated_by: voxislive-markdown-pipeline
---

[Home](https://voxislive.com/) · / · [Features](https://voxislive.com/features) · / · Session summary

SESSION SUMMARY

# A summary of the session — written when you ask for it, not before.

When a session ends, VoxisLive has already saved the transcript on your machine. The summary is a separate, deliberate step: you open the saved session, press one button, and a text recap comes back in the same language the transcript is written in.

## What the button does

Every session you run is saved as a transcript: the translated lines, and where available the source lines beside them. The summary feature reads that saved record and asks a cloud text model to condense it. It is a **post-session** action — there is no running commentary, nothing is summarised while you are listening, and nothing happens at all until you press the button.

The summary comes back in **the transcript's own language**. If you listened to a Spanish webinar in German, the transcript is German and so is the recap. There is no separate language picker for the summary, because there is nothing to pick: the record decides.

- Open a saved session from your history.
- Press the summary button.
- The recap appears as text you can read, select and copy.

This is a **paid-tier** feature and a manual one. It is not part of the free daily minutes, and it never fires on its own at the end of a session — see [pricing](https://voxislive.com/pricing) for what the paid tier includes.

## A meeting record is summarised differently

A one-way session — a video, a stream, a lecture — is a single voice going in one direction, and a plain recap is the right shape for it. A two-way [meeting](https://voxislive.com/use-cases/meetings) is not. It has sides, it has things that were agreed, and it has work that someone is supposed to do afterwards. So the meeting records are handed to the model already tagged.

Each line in a meeting transcript carries an **[mm:ss] offset** from the start of the session, a **Me / Them** side label, and — when more than one voice was detected on the other side — a speaker tag from VoxisLive's on-device **speaker-change detection**. That is change detection, not identification: it marks where the voice switched, it does not know who anyone is and never attaches a name.

With those tags in place, the meeting summary is asked for sections rather than a paragraph:

- **Summary** — what the conversation was about.
- **Decisions** — what was actually agreed, as opposed to discussed.
- **Action items** — one per line, each with a side against it: Me or Them.

A section the transcript has nothing for is left out rather than filled with a placeholder. A meeting where nothing was decided gets no Decisions heading — an empty section that reads as a finding is worse than no section.

The recap is written from the transcript and nothing else. It cannot recover a sentence the engine never captured, it cannot hear tone, and it cannot tell you what someone meant. If a line is missing from the transcript, it is missing from the summary.

## Fix the transcript first, then summarise

Live translation is not clean. A name comes through wrong, a number is mangled, a term the engine has never met is rendered as something close but not right. That is the honest state of the technology, and the useful answer is not to pretend otherwise — it is to let you correct the record.

Transcript lines are **editable in the app**. You can fix the translated text and the source line beside it, in place, on a saved session. Timing and the side and speaker tags stay as they were recorded; only the words change.

The order that matters: **correct first, then press the button**. The summary reads the saved record, so a fix you make before summarising is a fix the model sees. Corrections also flow into the exports, because the exports read the same saved record — fix a name once and it is right in the file you send round afterwards. More on that on the [transcript and history](https://voxislive.com/features/transcript-and-history) page.

## What it is not

The limits here are deliberate, and it is easier to say them plainly than to let you discover them:

| Limit | What that means in practice |
| --- | --- |
| Paid tier only | The summary is not part of the free daily minutes. |
| A button, never automatic | No summary is generated at the end of a session unless you ask for one. One press, one call. |
| No DOCX, no PDF | Transcripts export as TXT, SRT and VTT. There is no document export, and the summary is text you copy out of the app. |
| Long recordings are clipped | Past a certain length the whole transcript does not fit in one request. The head and the tail are kept and the middle is dropped — the opening frames what the session was about, the close carries the decisions. |
| No minutes template | It is a recap with sections, not a formal minutes format, and it is not a legal record of what was said. |

The clipping rule is worth understanding rather than working around. On a long session the summary has read the beginning and the end of your transcript, not the whole of it. If something decisive happened in the middle of a three-hour call, read that stretch yourself — the full transcript is still there, unclipped, in your history.

## Privacy: this button is the exception

Everything else about a transcript stays on your machine. The record is written to a folder in your own Documents, the history search runs locally, and the exports are files on your disk.

Pressing the summary button **sends the transcript text to a cloud model**. That is the one moment a saved transcript leaves the device, and it happens only when you press it. It is the reason the feature is manual: an automatic end-of-session summary would ship every transcript you ever record to a model without a decision from you, and that is not a decision we are willing to make on your behalf.

- Nothing is sent until the button is pressed.
- What is sent is the transcript text of that one session — not audio, and not your other sessions.
- If a session is confidential enough that its text should not leave the machine, do not press the button. The transcript, the editing and the TXT/SRT/VTT exports all work without it.

We hold no compliance certification and claim none. What we can tell you is the mechanism: which button sends what, where, and when. Judge it against your own policy rather than against a badge.

## Where it fits

The summary is the last step of a chain, not a feature on its own. Live translation gives you the session; the transcript gives you the record; editing gives you a record that is actually correct; the summary turns that corrected record into something short enough to send to somebody who was not there.

If that chain is what you are after, the [meetings use case](https://voxislive.com/use-cases/meetings) describes how a two-way call is captured in the first place, and [pricing](https://voxislive.com/pricing) covers the tier the summary sits on. The rest of the capability list is on [features](https://voxislive.com/features).

FAQ

## Common questions

### Is the summary generated automatically when a session ends?

No. It runs only when you press the button on a saved session. Nothing is summarised while a session is live, and nothing is generated in the background afterwards.

### What language is the summary written in?

The transcript's own language. If you listened to a call in French and the transcript is English, the summary is English. There is no separate setting for it.

### Can I get the summary as a Word document or a PDF?

No. There is no DOCX or PDF export. Transcripts export as TXT, SRT and VTT, and the summary is text in the app that you can select and copy.

### Does the summary work on the free tier?

No — it is a paid-tier feature. The free daily minutes cover live translation, not the summary call.

### Does pressing the button send my transcript anywhere?

Yes, and this is the one place a transcript leaves your machine. The transcript text of that session is sent to a cloud model to produce the recap. It happens only on the press; if you never press the button, the record stays on your device.

### I corrected a mistranslated name. Will the summary and the exported file use the correction?

Yes, provided you correct it first. Transcript lines are editable in the app, and both the summary and the TXT/SRT/VTT exports read the saved record — so a fix made before you summarise or export shows up in both.

### What happens on a very long recording?

Past a certain length the transcript does not fit in one request, so the head and the tail are kept and the middle is left out. The full transcript itself is untouched in your history — only what the model reads is clipped.


## Structured data (JSON-LD)

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "BreadcrumbList",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "name": "Home",
          "item": "https://voxislive.com/"
        },
        {
          "@type": "ListItem",
          "position": 2,
          "name": "Features",
          "item": "https://voxislive.com/features"
        },
        {
          "@type": "ListItem",
          "position": 3,
          "name": "Session summary",
          "item": "https://voxislive.com/features/session-summary"
        }
      ]
    },
    {
      "@type": "WebPage",
      "name": "Session summary: one button, paid tier | VoxisLive",
      "url": "https://voxislive.com/features/session-summary",
      "description": "A recap of a finished session, written when you press the button. Paid tier, never automatic, in the transcript's own language.",
      "inLanguage": "en",
      "isPartOf": {
        "@type": "WebSite",
        "name": "VoxisLive",
        "url": "https://voxislive.com/"
      },
      "publisher": {
        "@type": "Organization",
        "name": "VoxisLive",
        "url": "https://voxislive.com/"
      }
    },
    {
      "@type": "FAQPage",
      "mainEntity": [
        {
          "@type": "Question",
          "name": "Is the summary generated automatically when a session ends?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "No. It runs only when you press the button on a saved session. Nothing is summarised while a session is live, and nothing is generated in the background afterwards."
          }
        },
        {
          "@type": "Question",
          "name": "What language is the summary written in?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "The transcript's own language. If you listened to a call in French and the transcript is English, the summary is English. There is no separate setting for it."
          }
        },
        {
          "@type": "Question",
          "name": "Can I get the summary as a Word document or a PDF?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "No. There is no DOCX or PDF export. Transcripts export as TXT, SRT and VTT, and the summary is text in the app that you can select and copy."
          }
        },
        {
          "@type": "Question",
          "name": "Does the summary work on the free tier?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "No — it is a paid-tier feature. The free daily minutes cover live translation, not the summary call."
          }
        },
        {
          "@type": "Question",
          "name": "Does pressing the button send my transcript anywhere?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes, and this is the one place a transcript leaves your machine. The transcript text of that session is sent to a cloud model to produce the recap. It happens only on the press; if you never press the button, the record stays on your device."
          }
        },
        {
          "@type": "Question",
          "name": "I corrected a mistranslated name. Will the summary and the exported file use the correction?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes, provided you correct it first. Transcript lines are editable in the app, and both the summary and the TXT/SRT/VTT exports read the saved record — so a fix made before you summarise or export shows up in both."
          }
        },
        {
          "@type": "Question",
          "name": "What happens on a very long recording?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Past a certain length the transcript does not fit in one request, so the head and the tail are kept and the middle is left out. The full transcript itself is untouched in your history — only what the model reads is clipped."
          }
        }
      ]
    }
  ]
}
```