If you find that only part of your video has subtitles or a translation (for example, the opening seconds have no subtitles, a few sentences in the middle are missing, or some sections were not translated into the language you chose), this is not necessarily related to the free-plan limit. This article explains the common causes and how to fix them yourself.
First, check whether it's the free-plan limit
On the Free plan, the file-based modes (subtitle generation, transcript, AI meeting summary) only process the first 15 minutes of a file; the rest is not processed. The file itself uploads in full, but subtitles/transcript/summary are only generated for the first 15 minutes.
So the missing content is caused by the free-plan limit only if your content stops at around the 15-minute mark and you are on the Free plan. In the following cases, the cause is not the free-plan limit — see the troubleshooting steps below:
The video is shorter than 15 minutes in total, yet content is still missing;
The missing part is at the beginning (for example, the first 30 seconds have no subtitles) rather than the end;
You are already on a paid plan, but the content is still incomplete;
Individual sentences or passages are missing.
If what you are seeing is a translation that comes out in a language you didn't choose (for example, you selected Simplified Chinese but got English), that is a separate issue from the "missing content" cases above and is handled differently — please contact us using the details at the end of this article and include the affected project, and we will look into it for you.
(Note: the free trial for live caption translation is one-time, capped at 5 minutes — it does not renew — and is counted separately from the 15-minute limit for the file-based modes.)
How to get the rest of the content beyond the free 15 minutes
If you have confirmed that your content stops at the free 15-minute mark, there is no need to re-upload the file:
First, upgrade to any paid plan.
Open the project in the editor. At the point where the content cuts off, you will see a "Continue generating transcript" option (for transcript / AI meeting summary projects) or "Continue generating subtitles" (for subtitle projects).
The system will regenerate the remaining part of the same uploaded file (from the 15-minute cutoff to the end) and merge it into the existing project. This runs in the background.
Continuing generation deducts transcription minutes for the remaining content it generates; before you confirm, the screen shows what this operation will deduct.
What causes a missing opening or dropped sentences
When the missing content is not related to the free-plan limit, the most common causes have to do with the audio itself rather than a system error. The following situations are more likely to cause gaps:
The opening is silent or contains no speech — for example, the visuals-only intros common in vlogs, product videos, and beauty videos;
Music drowning out the voice — background music that is loud relative to the speech;
Heavy background noise — wind, crowds, or ambient noise;
A strong accent, or speech that is very quiet.
Even studio-quality, crystal-clear audio can still have gaps due to long pauses between sentences, technical terminology, very fast speech, or switching between multiple languages.
Re-transcribe it yourself (try this first)
For an affected video, you can trigger a re-transcription from inside the editor. It uses settings better suited to that video, rather than simply retrying with the same settings:
In the editor for the affected video, open the rating prompt (it pops up automatically, or you can open it via the button).
Choose "Not ideal".
Then choose "Subtitles are missing a lot of content".
The system will re-transcribe the video once: it keeps your existing subtitle track and creates a new track processed with different settings. This does not use any of your minutes.
You can run these steps right away for each affected video — no need to wait for us to handle them one by one. In most cases this improves the result, but it is probabilistic and may not fully resolve every case.
Consider the "High accuracy" transcription engine
If your video is one of the types above that are prone to gaps, you can choose the "High accuracy" transcription engine when creating the project. Compared with the standard engine, "High accuracy" recognizes words more accurately and is less likely to drop passages. This engine counts minute usage at 1.5 times the video length (video length × 1.5), and it is better suited to these kinds of videos — saving you the time of fixing lines by hand.
Still not resolved?
If the problem persists after re-transcribing, or if your translation is coming out in the wrong language, contact us at hello@subanana.com or via the in-app live chat, and include the details of the affected project — we will look into it further for you.