Frequently asked questions

Taption is a browser-based AI platform that turns video and audio into editable, time-coded text. Upload a file or paste a link, and Taption automatically produces a transcript, subtitles, translations into 50+ languages, speaker labels and AI summaries — with nothing to install.

Common uses include meeting minutes, interview transcripts, course and lecture captions, podcast show notes, and localised subtitles for YouTube and social video. New accounts get 20 free minutes, no credit card required.

Sign in and follow four steps:

• Click Upload.
• Choose a file from your computer, or paste a YouTube, Google Drive or other media link.
• Select the language that is actually spoken in the recording — this choice drives accuracy.
• Start the transcription.

Most files finish within a few minutes. You get a time-coded transcript you can play back, search and edit word by word. Minutes are deducted 1:1 against the length of the file. See how audio to text works.

Transcribe the video first, then open the subtitle view. Taption automatically splits the transcript into caption lines with timecodes, so you are correcting captions rather than creating them.

From there you can set a maximum number of characters per line, extend all subtitle durations at once, and smooth transitions so captions do not flicker between short pauses. Export SRT or VTT for YouTube and Vimeo, or export an MP4 with burned-in captions where you control font, size, colour, position and background. See the full captioning workflow.

Open a transcribed project, choose Translate and pick a target language from more than 50 options. Taption reuses the original timecodes, so translated subtitles stay in sync with the video automatically.

You can review the source and the translation side by side, edit either one, and then export a translated or bilingual SRT — or burn dual-language subtitles directly into an MP4. Translation deducts minutes 1:1 against the length of the file. See how video translation works.

Taption transcribes 40+ languages and translates into 50+ languages.

Transcription includes English (US, UK, Australia), Mandarin (Traditional and Simplified output), Cantonese, Japanese, Korean, Spanish, French, German, Italian, Portuguese, Dutch, Hindi, Vietnamese, Arabic and many more. Cantonese audio can be output as Traditional or Simplified Chinese.

The full list appears in the language selector when you upload. Choose the language spoken in the recording — selecting the wrong one is the single most common cause of a poor transcript.

Accuracy is typically 90–98% on clear recordings. The result depends on audio quality, background noise, accents, how often speakers talk over each other, and how much specialised terminology the recording contains.

Clean single-speaker audio recorded close to a microphone lands at the top of that range; a noisy multi-speaker recording full of jargon lands lower. Every transcript is fully editable, so the remaining corrections are made in the editor rather than by retyping from scratch.

Most of the gain comes from the recording itself, and the rest from a few settings:

• Record close to the microphone and keep background noise down.
• Ask people not to talk over one another — overlapping speech is the hardest case for any speech model.
• Select the correct spoken language, and the correct regional variant, when you upload.
• Add names, brands, acronyms and technical terms to your Personal AI Dictionary before transcribing.
• Use global search-and-replace to fix a repeated error in one pass.
• Use Translation Memory to keep terminology consistent across files.

The Personal AI Dictionary is a custom vocabulary list attached to your account. It holds the words a general-purpose speech model reliably gets wrong: people's names, company and product names, industry jargon and acronyms.

You add the correct spelling along with the way it tends to be mis-transcribed, and Taption applies the dictionary while processing your file. Because the list is stored on your account, a term you add once applies to every future upload — which makes it especially valuable for recurring content such as a weekly podcast, a course series or a project team's meetings.

Yes. Taption performs speaker diarisation: it detects when the speaker changes and splits the transcript into labelled segments.

Rename a generic label such as "Speaker 1" to a real name once, and the change is applied throughout the transcript. Speaker labels carry through to exported transcripts (TXT and PDF) and to subtitles, which makes them useful for interviews, panel discussions, meeting minutes and research recordings. Results are best when speakers do not overlap heavily. See how speaker identification works.

Yes — everything the AI produces is editable. In the editor you can:

• Click any word to jump the video to that exact moment.
• Correct text inline, with the video staying in sync.
• Merge and split caption lines, and drag timecodes on the timeline.
• Run a global search-and-replace across the whole transcript.
• Set a maximum number of characters per caption line.
• Extend every subtitle's duration at once, or smooth transitions between them.

Changes save as you work and flow into every export you generate afterwards.

Yes. Taption supports text-based video editing: delete a sentence in the transcript and the matching segment is removed from the video, so you trim by selecting text instead of scrubbing a timeline.

Subtitles and timecodes re-adjust automatically after each cut, which means exported transcripts, subtitle files and burned-in videos all stay in sync. It is the fastest way to strip out filler, false starts and off-topic tangents before publishing.

Yes. Upload a video or paste a YouTube link and run AI Analysis. Ready-made templates cover the common jobs, so you do not have to write prompts.

Chat with Video lets you ask questions such as "what were the action items?" or "what was decided about the budget?" and get answers grounded in the recording rather than in a general model. Analysis is available in 50+ languages.

AI Analysis uses AI credits: Premium includes 5 per month and Bulk includes 30 per month, and more can be purchased at any time. See what AI Analysis produces.

Yes. The AI Analysis templates produce:

• A concise summary of the whole recording.
• Bulleted key takeaways.
• Time-coded YouTube chapters you can paste straight into a video description.
• Meeting notes with decisions and action items.
• Title and outline suggestions.
• A cleaned-up, readable version of the raw transcript.

Every result is editable and copyable, so you can adjust the wording before publishing or sharing it.

Video: MP4, MOV, AVI, WMV, FLV, MPEG, MPG and OGG.
Audio: MP3, WAV, M4A, AAC, FLAC and OGG.

Taption handles compatibility for you, so there is no need to convert files manually first.

A single file can be up to 5GB and up to 5 hours long. The 5GB limit is measured against the MP4 equivalent, so a large file in another format can still exceed it once converted.

Yes. In the upload dialog, paste the YouTube URL instead of choosing a file. Taption pulls the audio and transcribes it.

From there you can translate it, generate subtitles and export SRT or VTT to upload back to YouTube, or generate time-coded chapters for the description. The same works for other public media links, including TikTok, Instagram, Facebook, Vimeo and direct media URLs.

One caveat: a video uploaded to YouTube within roughly the last 24 hours may not be importable yet.

Yes. Click Upload, choose Zoom or Webex, sign in and authorise the connection, then pick a recording from the list and transcribe it — no downloading and re-uploading. Google Drive, Dropbox and OneDrive links are supported the same way.

Step-by-step guides: Zoom integration and Webex integration. You can disconnect an integration at any time under Account → Apps → Uninstall.

• Subtitles: SRT and VTT.
• Transcripts: TXT and PDF, with speaker labels and timestamps, using export templates you can customise.
• Editing timelines: FCPXML for Final Cut Pro, as editable text layers or a standard caption track.
• Video: MP4 with burned-in subtitles, including dual-language subtitles, at up to 4K original quality.

Audio-only files can also be exported as an MP4 video with a custom background image — useful for turning a podcast episode into something publishable on YouTube. See audio to video.

Minutes are the unit of usage, and they are deducted by the length of your media, not by how long processing takes:

• Speech-to-text: 1:1 against the file length.
• Translation: 1:1 against the file length.
• Burned-in MP4 export: 1:1 for pay-as-you-go, free and unlimited on Premium and Bulk subscriptions.

So transcribing a 30-minute video and then translating it uses 60 minutes in total. New accounts start with 20 free minutes.

Subscription minutes roll over each month while the subscription is active, and are always used before separately purchased minutes. Add-on minutes you buy never expire, even if you later cancel.

You will see Not enough minutes or Exceed trial usage when you try to upload.

Nothing is lost. Your existing transcripts stay in your account and the editor keeps working — you can still edit, translate what is already processed, and export. To continue transcribing, purchase add-on minutes or subscribe, then retry the upload.

Yes, at any time — and separately purchased minutes never expire, even if you later cancel a subscription. The rate depends on your plan:

• Pay as you go: $8/hour
• Premium Subscription: $6/hour
• Bulk Subscription: $4/hour

Subscriptions also include monthly minutes — 120 with Premium and 1000 with Bulk — which roll over while the subscription stays active. Compare plans and pricing.

Once you have subscribed or purchased minutes, go to Account → Invoices → View All. Every past invoice can be viewed and downloaded as a PDF there.

Bulk subscribers can issue invoices carrying a custom company or institution name and a tax ID number. For wire transfer, cheque or other custom invoicing arrangements, email help@taption.com.

Yes. Uploads and downloads are protected with HTTPS encryption, and your audio, video and transcripts are processed only to produce the transcription, subtitles, translations and AI summaries you request.

You stay in control of your own content: files and transcripts can be deleted from your account at any time, and team projects are governed by role-based permissions, so only the members you invite can view, edit or export. See the privacy policy for how data is collected, stored and removed.

Taption already has a range of information security and data protection measures in place, and continues to raise its enterprise security and compliance capability. Current technical measures include HTTPS encryption in transit, role-based permissions and user-controlled deletion of data.

As Taption has grown and enterprise customers increasingly run formal security and compliance reviews, we are using Vanta to further formalise our security policies, controls and evidence management, with applicable technical controls under automated continuous monitoring.

Building on the security measures already in place, we are currently advancing:

• SOC 2 Type II audit readiness
• ISO 27001 certification readiness
• HIPAA security and privacy compliance readiness
• Independent third-party manual black-box penetration testing

The third-party penetration test will have an external team verify the effectiveness of existing safeguards in practice. We will also progressively provide enterprise customers with our security policies, controls, compliance status and third-party validation information through a Trust Center.

For any security or compliance question, email us at help@taption.com.

By far the most common cause is the wrong source language being selected at upload. Translation runs from the transcript, so if the spoken language was set incorrectly the transcript itself is wrong, and the translation simply inherits the problem.

Open the project and check which language it was transcribed in. If it is wrong, re-transcribe with the correct spoken language selected. Also confirm you have enough minutes left, since translation deducts 1:1 against the length of the file.

Work through this checklist:

• Size and length: the file must be under 5GB and under 5 hours.
• Video size too large on a file under 5GB: the limit is measured against the MP4 equivalent, and converting from another format can push it over. Re-encode to MP4 first.
• Format: confirm it is on the supported list above.
• Audio track: a video with no audio has nothing to transcribe.
• Minutes: make sure your balance covers the file length.
• Connection: large uploads need a stable one.

If it still fails, email help@taption.com with the file name and the exact error message.

Youtube link failed, please try again almost always means the video was uploaded very recently. YouTube needs time to make a new upload fully available, so wait at least 24 hours and try again.

Otherwise, check that the video is public or unlisted — private, age-restricted and region-blocked videos cannot be imported — and that you pasted the full watch URL. If you need the transcript immediately, download the video and upload the file directly instead.

Email help@taption.com for anything to do with your account, billing and invoices, integrations, or a specific file that is not processing correctly. Including your account email and the project name or exact error message gets you a faster answer.

The same address handles enterprise plans, custom API integration, the affiliate programme, and onboarding or training for teams. See team and collaboration features.