# Sundra
> Sundra turns the materials you already have into structured training courses in hours, then delivers them in 70+ languages. Built for HR and L&D teams.
# Downloading Text Transcripts and Scripts From Your Videos
URL: https://www.sundra.io/guides/how-to-download-text-transcripts-and-scripts
Last updated: 2026-10-07
Sometimes you need the words without the video: a printable handout for a training session, a script to review with a subject-matter expert, a text version for an accessibility request, or source material for a course page. Every captioned video in Sundra already has a full transcript. Here are two ways to get it out.
- **Download a TXT file** from the Export dialog: best for a clean, complete transcript, and for several languages at once.
- **Ask the Assistant** for the transcript: best when you just want to copy the text into another document.
## Before you begin
You'll need a video in Sundra that has **finished captioning**. For transcripts in other languages, translate the captions first. Every translation shows up as an export option.
## Option 1: Download a TXT file
### Step 1: Open the Export dialog
Open your video from the **Library**. In the **Captions** panel on the right, click **Export**. The **Export Project** dialog opens on the **Captions** tab, which is the one you want.
### Step 2: Pick your languages
Leave **Export To** set to **Download File**. Then open the **Languages** dropdown and pick the language, or languages, you want a transcript in. It lists the source language and every translation of the video.
### Step 3: Choose Text file (.txt)
Open the **Format** dropdown and choose **Text file (.txt)**. That gives you the spoken text only (no timecodes or numbering), ready to read, print or paste.
Need subtitles instead? The same menu offers **SubRip titles (.srt)** and **Video Text Tracks (.vtt)**, the caption formats most video players and LMSs accept.
### Step 4: Click Export
Click **Export**, and the transcript downloads to your computer. Pick several languages and you get a single ZIP file with one transcript per language.
## Option 2: Ask the Assistant
### Step 1: Open the Assistant
In the panel on the right of the editor, click the **Assistant** tab.
### Step 2: Ask for the transcript
Type a request into the message box, such as _"Could I get the transcript?"_, and press **Enter**. The Assistant reads the video's captions and writes the transcript out as text.
### Step 3: Copy the text
The complete transcript appears in the panel. Select it and copy it into your document.
The Assistant can also reshape the text for you: ask for _"the transcript as a script with speaker notes"_, _"the transcript in German"_ or _"the transcript split into paragraphs"_.
## Good to know
- **The transcript matches your captions.** Any fixes you make in the caption editor show up in the next export, so correct names and terms before you download.
- **TXT for reading, SRT/VTT for playing.** Use a TXT file for documents and handouts; use SRT or VTT when the captions need to play in time with the video.
- **Want summaries rather than the full text?** See [Generating text from a video](/guides/how-to-generate-text-from-video).
## That's it!
Your transcript is ready to share, print or build on. Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Exporting a Video With Burnt-In Subtitles
URL: https://www.sundra.io/guides/how-to-export-videos-with-burnt-in-subtitles
Last updated: 2026-10-07
Once your captions are ready, you'll often want a video file with the subtitles **burned in**, drawn directly onto the picture, so they show up everywhere the video plays: in your LMS, on an intranet page, in a social post, or on a screen in the break room. No separate caption file, no player settings to get right.
This guide walks you through exporting that video from Sundra, from the editor to a downloaded MP4.
## Before you begin
You'll need a video in Sundra that has **finished captioning**. If you want subtitles in other languages, translate the captions first. Every translation you've made shows up as an export option.
## Step 1: Open your video and click Export
Open your video from the **Library**. In the editor, the **Captions** panel on the right has an **Export** button at the top. Click it.
## Step 2: Choose the Video tab
The **Export Project** dialog lets you export either **Captions** (subtitle files such as SRT or VTT) or a **Video** with the captions embedded. Select the **Video** tab.
Leave **Export To** set to **Download File** to get an MP4 you can save.
## Step 3: Pick your caption languages
Open the **Caption Languages** dropdown and pick the languages you want burned in. It starts at **No captions (original video)**, and lists your source language plus every translation of the video.
You can select more than one language: Sundra renders **one video per language**, so you get a separate MP4 for each.
## Step 4: Click Export
Click **Export**. The render runs in the background, so you can close the dialog and keep working (edit another video, translate, or grab a coffee).
## Step 5: Follow the progress in the Exports widget
Click the **download icon** in the top bar to open the **Exports** widget. Each video being rendered appears there with its language and a progress bar. A badge on the icon tells you how many exports are running, even when the widget is closed.
## Step 6: Download your MP4
When the render is done, a **download** button appears next to the video in the Exports widget. Click it, and your MP4, with the subtitles burned in, is saved to your computer.
## Tips
- **Style before you export.** Subtitles are burned in exactly as they look in the editor preview, so set the font, size and colours under **Text Appearance** in the Captions panel first.
- **Caption fixes need a new export.** Burned-in subtitles are part of the picture, so if you edit the captions afterwards, export the video again.
- **Want captions you can turn off?** Use the **Captions** tab in the same dialog to download a subtitle file instead, and upload it alongside the original video.
## That's it!
You now have a video with subtitles that play everywhere, in every language you need. If anything doesn't look right, reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Finding and Downloading Your Previous Video Exports
URL: https://www.sundra.io/guides/how-to-find-previous-video-exports
Last updated: 2026-10-07
When you [export a video with burnt-in subtitles](/guides/how-to-export-videos-with-burnt-in-subtitles), Sundra keeps the finished MP4. So if you need the file again (on another computer, for a colleague, or because the download went to the wrong folder), there's no need to render it a second time.
There are two places to find your exports.
## Option 1: From the video's Export dialog
This is where every completed export of a video lives, however long ago you made it.
### Step 1: Click Export
Open your video from the **Library**, and click **Export** at the top of the **Captions** panel on the right.
### Step 2: Select the Video tab
In the **Export Project** dialog, select the **Video** tab. Below **Export To**, a table lists every language this video has been exported in, with its **Status**.
### Step 3: Download the one you need
Click the **download** icon next to any export marked **Complete**, and the MP4 is saved to your computer.
## Option 2: From the Exports widget
The **download** icon in the top bar opens the **Exports** widget. It shows exports that are still rendering and the ones that finished recently, across all your videos, each with its language and its own **download** button. A badge on the icon tells you when something is in progress.
It's the quickest way to grab a render you've just started. For anything older, use the Export dialog in Option 1.
## Good to know
- **Edited the captions since?** A previous export still has the old subtitles burned in. Export the video again to get your latest changes.
- **Need a language that isn't listed?** It hasn't been exported yet. Pick it under **Caption Languages** in the same dialog and click **Export**.
## That's it!
Your finished videos are always a couple of clicks away. Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Generating Text From a Video With the Sundra Assistant
URL: https://www.sundra.io/guides/how-to-generate-text-from-video
Last updated: 2026-10-07
A training session, an all-hands or an expert interview holds far more than the video itself: the key points, the decisions, the quotes worth sharing. Once Sundra has captioned a video, the **Assistant** can turn what was said into text you can use right away: key notes, a summary, meeting minutes or a first draft of a blog post.
This guide shows you how to ask for it, from the caption editor to the finished text.
## Before you begin
You'll need a video in Sundra that has **finished captioning**. The Assistant works from the captions, so the text it writes is only as accurate as they are. Fix any misheard names first (or add them as [Transcription Keyterms](/guides/how-to-use-transcription-keyterms) before you upload).
## Step 1: Open the Assistant
Open your video from the **Library**. In the panel on the right of the editor, click the **Assistant** tab, next to **Captions** and **Activity**.
## Step 2: Ask for the text you need
Click into the **Write anything…** box at the bottom of the panel and describe what you want, in your own words. For example:
- _"Summarize this video into key notes"_
- _"Write meeting minutes with the decisions and action items"_
- _"Draft a short blog post based on this talk"_
- _"List the three most important points for a new employee"_
The more specific you are about length, tone and audience, the closer the first draft will be.
## Step 3: Send it and let the Assistant work
Press **Enter** (or click the send button) to generate your text. The Assistant reads the video's captions (you'll see **Reading captions** while it does) and then writes its answer.
## Step 4: Copy your text
The answer appears in the panel, formatted with headings and bullet points. Select the text and copy it into your document, email, LMS or intranet page.
Not quite right? Just reply in the same conversation (_"make it shorter"_, _"add a section on safety"_, _"write it for managers"_) and the Assistant revises it.
## Ideas for what to ask
- **Key notes and summaries** for people who missed the session
- **Meeting minutes** with decisions and action items
- **Blog post or newsletter drafts** from a talk or interview
- **Quiz questions** to check understanding after a training video
- **Caption fixes and translations**, straight from the same panel
## Good to know
- **The Assistant answers about the video you have open.** Open a different video to ask about it.
- **Check the facts.** The Assistant works from what was said in the video, but always read generated text through before you publish it.
- **Need the word-for-word transcript instead?** See [Downloading text transcripts and scripts](/guides/how-to-download-text-transcripts-and-scripts).
## That's it!
Every captioned video is now a source of ready-to-use text. Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Styling Captions and Setting the Line Width in Sundra
URL: https://www.sundra.io/guides/how-to-style-captions-and-line-width
Last updated: 2026-10-07
Good captions are easy to read at a glance: lines that aren't too long, a font that suits your brand, and enough contrast to stand out against any background. In Sundra, all of that lives in a **caption style**: a set of appearance settings saved to your workspace, so every video can look the same.
The most important setting is the **line width**. Instead of pressing Enter to break long captions by hand, you tell Sundra how many characters fit on a line, and it wraps every caption for you.
## Before you begin
Open a captioned video from the **Library**. Everything in this guide happens in the **Captions** panel on the right of the editor, and the player shows your changes as you make them.
## Step 1: Open Text Appearance
In the **Captions** panel, click **Text Appearance** to expand it. It shows the style the video is using (**Default** to begin with).
## Step 2: Create a new style
Click **Create New Style**.
## Step 3: Name your style
Give the style a name you'll recognise later (for example, _Course captions_ or _Social clips_) and click **Create Style**. The new style is selected for this video straight away.
## Step 4: Set the line width
**Line width** is the maximum number of characters on one line of a caption. Use the **−** and **+** buttons, or type a number. Any caption longer than that breaks onto a new line automatically, at a space between words.
Around **35–42 characters** reads comfortably on most screens; go lower for videos that will be watched on phones or in vertical formats.
## Step 5: Adjust the appearance
Fine-tune the rest of the style. Every change shows in the player straight away:
- **Font and size**: pick any font from the list and a point size.
- **Text color**: with an opacity percentage.
- **Stroke**: an outline around the letters, with a colour and thickness in pixels.
- **Background**: a box behind the text, with a colour and opacity. A semi-transparent dark background keeps captions readable over busy footage.
- **Alignment**: left, centre or right.
- **Position**: top, middle or bottom of the frame.
- **Casing**: as typed, ALL CAPS, lowercase or Title Case.
- **Karaoke**: highlights each word, in a colour you choose, as it is spoken.
## Step 6: You're done, and changes save automatically
There's no save button: your changes are stored in the style as you make them. Because styles belong to your **workspace**, you and your teammates can pick the same style from the dropdown on any other video.
## Good to know
- **What you see is what you export.** Burned-in subtitles are rendered exactly as they look in the player, so set the style before you [export a video with burnt-in subtitles](/guides/how-to-export-videos-with-burnt-in-subtitles).
- **Line width only changes how captions wrap.** Your caption text and timing stay as they are.
- **One style per purpose.** Keep separate styles for different outputs: for example, a larger, narrower style for phone-first social clips and a wider one for course videos.
## That's it!
Your captions now break cleanly and match your brand on every video. Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Uploading a Video and Getting Captions in Sundra
URL: https://www.sundra.io/guides/how-to-upload-and-caption-videos
Last updated: 2026-10-07
Every captioned video in Sundra starts with an upload. You add the file from your **Dashboard** (or the **Library**), Sundra auto-detects the language spoken in it (you can also pick it yourself), and transcribes it into captions you can edit, translate and export.
This guide walks through an upload from start to finish, and covers what to do if an upload doesn't go through.
## Supported formats
Sundra takes the common video formats:
- **MP4** and **MOV**, the formats most cameras, phones and editing tools export
- **M4V**, **MKV** and **WebM**
Each file can be up to **5 GB**, and you can add up to **20 files** in one go. If you're exporting from an editor, an **MP4 (H.264)** file is the safest choice.
## Step 1: Open the upload area on your Dashboard
On your **Dashboard**, click **Translate & caption video**. The upload area opens right below it.
You can also upload from the **Library**, which has the same upload area at the top.
## Step 2: Add your video
**Drag your video onto the page**, or click **"Drop a video anywhere on this page or click here to upload"** to pick it from your computer.
Have the video on Vimeo instead? Click **Import from Vimeo** (see [Connecting Vimeo to Sundra](/guides/how-to-connect-vimeo-to-sundra)).
The video appears in the upload area with its name, type, length and size. This is your chance to check you picked the right file. Remove it with the **×**, or click **Add more files** to upload several together.
## Step 3: Check the spoken language (optional)
By default the language is set to **Auto-detect**: Sundra listens to the audio and sets the spoken language for you once the transcription is done. You don't need to do anything here.
If you'd rather set it yourself, open the language menu next to the file and pick the language spoken in the video. Type in the search box to find it quickly.
> **Tip:** Picking the language yourself is the most reliable option when you know it, especially for short clips with little speech.
## Step 4: Click Upload
Click **Upload**, whether you kept **Auto-detect** or picked a language.
The video gets its own card under **Recent** straight away, with a progress bar showing how much has been sent.
**Keep the tab open until the upload reaches 100%.** You can keep working elsewhere in Sundra, but closing or reloading the tab stops the upload.
## Step 5: Let Sundra transcribe it
Once the upload finishes, Sundra prepares the video and starts transcribing it right away, and the card moves from **Waiting** to **Transcribing**. This part happens on our side, so you can close the tab now; you'll find the video waiting when you come back.
## Step 6: Open your captioned video
When the card says **Completed**, your captions are ready. Click the card (on the Dashboard or in the **Library**) to open the video in the caption editor, where you can fix any word, adjust timing, translate and export.
## Troubleshooting: "Upload failed"
If something goes wrong, the video's card turns red and tells you why. Here are the common reasons and what to do about each.
### "The upload was interrupted."
The connection dropped or the tab was closed before the upload finished. Delete the failed card, check your internet connection, and upload the file again, and keep the tab open until the progress bar reaches 100%. On a slow or unstable connection, try a wired network or upload outside busy hours.
### "This video file looks incomplete"
The file itself is cut off: a recording or screen capture stopped too early, an export never finished, or the file was copied before it was fully saved. Open the file on your computer and check it plays all the way to the end. If it doesn't, export or copy it again from the original and upload the new file.
### "The file could not be processed."
The file uploaded, but Sundra couldn't read the video inside it. This usually means an unusual codec or a damaged file. Re-export the video as **MP4 (H.264 video, AAC audio)** from your editing tool or a free converter such as HandBrake, then upload it again.
### The file was skipped before it uploaded
If the file never gets a card and you see a message like _"… was skipped"_, Sundra didn't accept it:
- **"we take video, audio, …"**: the format isn't supported. Convert it to MP4 or MOV.
- **"it is over 5 GB"**: the file is too large. Trim it, split it into parts, or export at a lower bitrate.
### The card is stuck on Uploading
If you closed the tab mid-upload, the card can't finish. Delete it and upload the file again.
Still stuck? Use the chat button in the app or email [hello@sundra.io](mailto:hello@sundra.io) with the file name and the message you see, and we'll take a look.
## Good to know
- **Auto-detect works best on clear speech.** If the detected language is wrong, re-upload the video and pick the language yourself.
- **Make names come out right.** Add company names, product terms and jargon as [Transcription Keyterms](/guides/how-to-use-transcription-keyterms) before you upload.
- **Uploading several videos?** Add them all at once, and set the language for each file before clicking **Upload**.
## That's it!
Your video is in the Library and captioned. Next, you might want to [style your captions](/guides/how-to-style-captions-and-line-width) or [export the video with burnt-in subtitles](/guides/how-to-export-videos-with-burnt-in-subtitles). Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Using Transcription Keyterms for More Accurate Transcripts
URL: https://www.sundra.io/guides/how-to-use-transcription-keyterms
Last updated: 2026-10-07
Sundra's AI transcription handles everyday speech very well. Where any speech-to-text model struggles is with words it has rarely or never heard: company and product names, people's names, acronyms, and industry jargon. In an interview or an expert talk, those are often the words that matter most, and a misheard name in the captions is the first thing your audience notices.
**Transcription Keyterms** fix that. It's a glossary for your workspace: list the words that are uncommon or hard to hear, and Sundra tells the AI to listen out for them every time it transcribes.
## What makes a good keyterm
Add words the AI is unlikely to guess on its own, for example:
- **Company, brand and product names**: e.g. _Össur_, _Marel_, _Carbfix_
- **People's names**, especially interviewees and speakers with uncommon spellings
- **Industry terms and acronyms**: medical, legal, engineering or internal jargon
- **Place names** and words borrowed from another language
There's no need to add everyday words. The AI already knows those, and a short, focused list works best.
## Step 1: Open Workspace Settings
Click **Settings** under **Workspace** in the left sidebar.
## Step 2: Find the Transcription Keyterms section
The **Transcription Keyterms** section is at the top of **Workspace Settings**. Keyterms belong to the workspace, so everyone on your team benefits from the same list.
## Step 3: Type your terms
Click into the **Add Keyterms** field and type a term, then press **Enter** or type a **comma** to stage it. Staged terms appear as chips, so you can add several in a row and remove any typo with its **×** before saving.
Each keyterm can be up to **5 words** and **50 characters**, and a workspace can hold up to **1,000** keyterms. The counter in the corner shows how many you've used.
## Step 4: Click Add to save them
Click the **Add keyterms** button (it shows how many terms are staged, e.g. **Add 4 keyterms**) to save the staged terms to your workspace. They move to the **Saved Keyterms** list below.
From here you can remove a single term with its **×**, or empty the list with **Clear all**.
## Step 5: Transcribe as usual
That's all the setup there is. Upload your interview the way you always do, and Sundra passes your saved keyterms to the AI along with the audio.
## Good to know
- **Keyterms apply to new transcriptions.** Videos transcribed before you added a term aren't changed. Upload the video again if you want it re-transcribed with the updated list.
- **Spell terms the way you want them to appear.** The AI uses your spelling, capitalisation and accents, so _Össur_ comes out as _Össur_, not _Ossur_.
- **Keep the list current.** When a project ends or a product is renamed, tidy up the list so it stays focused on the words you actually need.
## That's it!
With your keyterms in place, your transcripts start out closer to perfect, and you spend less time fixing names in the caption editor. Questions? Reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# Connecting Vimeo to Sundra
URL: https://www.sundra.io/guides/how-to-connect-vimeo-to-sundra
Last updated: 2026-05-22
Connecting your Vimeo account to Sundra lets you import and export videos and captions directly from your Vimeo library - no downloading and re-uploading required. This guide walks you through every step, from creating a Vimeo Developer app to linking it in Sundra.
## Before you begin
You'll need a **Vimeo account** (free or paid) and a **Sundra account**. The whole setup takes about five minutes.
## Step 1: Go to the Vimeo Developer Portal
Open your browser and navigate to **developer.vimeo.com**. Log in with the same Vimeo account that holds the videos you want to caption in Sundra.
## Step 2: Create a new app
Click **"Create app"** (or **"New App"**) in the top-right area of the developer dashboard. Fill in the required fields:
- **App Name**: something descriptive, e.g. "Sundra Integration"
- **App Description**: briefly explain the purpose, e.g. "Connect Vimeo to Sundra for captioning and translation"
- **Will People besides you be able to access your app?**: Select "No"
Accept the terms and conditions and click **"Create app"** to save. Vimeo will take you to your new app's settings page.
## Step 3: Generate a Personal Access Token (PAT)
On your app's settings page, scroll down to the **"Authentication"** section. This is where you create the token Sundra will use to access your videos.
Before generating the token, you need to select the right **scopes** (permissions). Make sure the following are checked:
- **Private**: allows access to your private videos
- **Video Files**: allows Sundra to download video files for captioning
- **Upload**: allows Sundra to upload captioned videos back to Vimeo
Click **"Generate"**. A long token string will appear. **Copy it immediately**. Vimeo only shows the full token once. If you lose it, you'll need to generate a new one.
## Step 4: Connect Vimeo in Sundra
Now head over to **app.sundra.io** and log in. In the sidebar under **Workspace**, click **Settings** to open Workspace Settings, then find the **"Connected Integrations"** section.
Click **Connect with Access Token**. In the popup, paste your Personal Access Token into the token field and click **"Connect"**. Sundra will verify the token and confirm the connection. You should see a success message and your Vimeo account name displayed.
## Step 5: Import videos from Vimeo
Once connected, go to your **Dashboard** in Sundra. When uploading a new video, you'll now see an option to **import from Vimeo**. Simply paste the link to the video you want to caption, and Sundra will pull it in automatically, with no manual download needed.
## Troubleshooting
- **Token not working?** Double-check that you selected the **Private**, **Video Files**, and **Upload** scopes when generating the token. If in doubt, generate a new one with the correct permissions.
- **Can't see your videos?** Make sure you're logged into the same Vimeo account that owns the videos. Team members on Vimeo Business or Premium plans may need the account owner to create the app and token.
- **Connection lost?** Tokens don't expire by default, but if you revoke or regenerate the token on Vimeo, you'll need to update it in Sundra Workspace Settings.
## That's it!
Your Vimeo account is now linked to Sundra. You can import videos, caption them, translate into 70+ languages, and export, all without leaving the platform. If you run into any issues, reach out to us at [hello@sundra.io](mailto:hello@sundra.io).
---
# We Asked 71 HR Pros to Spot the AI. On Voices, They Did No Better Than a Coin Flip.
URL: https://www.sundra.io/articles/ai-or-not-results
Last updated: 2026-10-06
On 2 October we had a booth at [Mannauðsdagurinn 2026](https://mannaudsfolk.is/myndir-fra-mannaudsdeginum-2026/), the annual HR Day held by Mannauður, the Icelandic association of HR professionals. At the booth we ran an arcade game called **AI or Not**. Each game had three rounds: four faces, five voice clips and five videos. Players had ten seconds per clip to decide whether it was real or AI-generated, and earned more points for fast answers and streaks of correct ones.
71 people played, making 1,092 guesses in total.
## Results by round
Players were right 60% of the time overall, but the rounds differed a lot:
- Video was the easiest round, at 71% correct.
- Faces came next at 60%. Most of that came from recognising real faces. Players identified fewer than half of the AI faces.
- Voices came last at 48%, which is no better than guessing. Players did worse on real human voices than on AI ones.
The overall figure for each round has a margin of error of about ±5 percentage points (95% confidence), so the gap between video and voices holds up. The real and AI figures within each round are less precise, at about ±7 points.
## Voices: real narrators were mistaken for AI
Players identified 53% of the AI voices correctly but only 44% of the real recordings.
The real clips were volunteer readings of classic books from LibriVox, including Alice in Wonderland, Moby Dick, Sherlock Holmes and The Art of War. On the English recordings, players recognised the human reader 39% of the time. All seven players who heard one Alice in Wonderland reading got it wrong. Listen to it next to an AI voice reading the same book:
Real: a LibriVox volunteer. 0 of 7 players got it right.
AI: an ElevenLabs voice. 3 of 11 players got it right.
One likely explanation is that people expect AI voices to sound smooth and even, so a calm, clearly recorded human reader sounds artificial to them. The AI voices, made with ElevenLabs, can include breaths, hesitations and small irregularities, which are the cues listeners use to decide a voice is human.
The voice round also included Icelandic clips. Players identified 41% of the AI-generated Icelandic voices (23 of 56 guesses), against 59% of the English AI voices. The Icelandic sample is small, so the difference may not hold up with more players.
## Faces: AI faces usually passed as real
Players recognised real faces 74% of the time but identified only 45% of the AI-generated faces.
When unsure, players tended to answer "real". Only about a third of all answers in the faces round were "AI". The hardest AI face in the game fooled 9 of the 11 players who saw it. Both of these faces are AI-generated:
AI-generated. 2 of 11 players got it right.
AI-generated. 4 of 17 players got it right.
The AI faces came from StyleGAN, an image generator released several years ago. Faces from current image models are more realistic.
## Video: the easiest round, depending on the clip
Players got 71% of video clips right, with similar results for real (70%) and AI (72%) clips. AI video often still has visible flaws, such as odd physics, distorted hands and unnatural camera movement.
Results varied with the type of clip:
| Video type | Real clips spotted | AI clips spotted |
| --- | --- | --- |
| Silent scenes (dancers, a barista, a city street) | 57% | 76% |
| People talking to camera | 76% | 64% |
Silent AI scenes were the easiest clips in the game to identify. More than a third of players took the AI clips of people talking to camera for real. We generated those clips with Google's Veo 3; some show AI "vloggers" speaking Icelandic.
Players also got real silent stock footage wrong 43% of the time. Nine of the 11 players who saw a real clip of a ballerina got it wrong.
AI: a Veo 3 "vlogger" speaking Icelandic. 6 of 11 players got it right.
Real: footage of a ballerina (Pexels). 2 of 11 players got it right.
## The hardest clips
Of the clips shown to at least seven players, these had the fewest correct answers:
| Clip | What it really was | Players right |
| --- | --- | --- |
| [Alice in Wonderland reading](/aiornot-media/m/43fd32a73f567d01.mp3) | Real human narrator | 0 of 7 |
| [Moby Dick reading](/aiornot-media/m/5d9d0d3a2a31daa4.mp3) | Real human narrator | 1 of 7 |
| [The Art of War reading](/aiornot-media/m/8a0f4b417bed6a06.mp3) | Real human narrator | 2 of 12 |
| [Portrait](/aiornot-media/m/8a5940eb33006cbc.jpg) | AI-generated face | 2 of 11 |
| [Ballerina](/aiornot-media/m/18e97d800093d273.mp4) | Real footage | 2 of 11 |
| [Portrait](/aiornot-media/m/36e95e716e68c68f.jpg) | AI-generated face | 4 of 17 |
Three of the six are recordings of real people reading aloud.
## Scores
Scores combined accuracy, speed and streaks. The median score was 1,368, half of all players scored between 1,163 and 1,610, and the top score was 2,645.
The scores form a rough bell curve that leans slightly left: most players sit a little below the middle, and a short tail of high scores pulls the average (1,393) just above the median. The streak bonus is a likely cause, since it multiplies points for every correct answer in a row, so a good run adds far more than a bad one takes away. With 71 players the lean is too small to be sure of, and no group of players stood out at the top.
## What this means for HR teams
**AI voices and presenters are ready for training material.** More than a third of players took our AI presenters for real people, and players spotted AI voices only about half the time. For training content, that means narration and on-screen presenters no longer have to be recorded in a studio.
**Icelandic is no longer a gap.** Players identified only 41% of the AI-generated Icelandic voices, fewer than the English ones. Training material in Icelandic used to mean either a human recording or a robotic voice. AI voices can now produce Icelandic narration that most listeners accept as natural.
**Label AI content clearly.** If employees can't tell an AI presenter from a real one, they shouldn't have to guess. Mark AI-generated presenters and voices in your training material, so people know what they're watching and trust the content because of what it says, not who appears to say it.
**Don't verify people by voice.** The same technology can be misused. A call from "the CEO" asking for an urgent payment, or a job candidate's video introduction, can be generated. Confirm requests like these through a second channel you already trust.
## How Sundra uses AI
Sundra uses the same kinds of AI models that fooled our players, to make training material easy to create and accessible for a diverse workforce.
Most companies already have the knowledge their employees need. It sits in handbooks, safety manuals, policies and meeting notes. Sundra turns that material into complete courses:
- **Course content:** AI organises the source material into modules and lessons, and writes the text for each one.
- **AI presenters:** each lesson can be presented on video by an AI presenter, so there's no filming, and no reshoot when a procedure changes.
- **Narration and languages:** courses are voiced with AI narration in more than 70 languages, so every employee can take them in their own language.
Subject matter experts review and approve every word in Sundra's editor before a course is published. Sundra is built with responsible use of AI in mind.
## Method and limitations
- **Players:** visitors to our booth at Mannauðsdagurinn 2026. They chose to play, so they aren't a representative sample.
- **Format:** each game dealt 4 faces, 5 voices and 5 videos at random from a pool of 126 clips. Each clip had a 10-second timer, and an answer that ran out of time counted as wrong.
- **Data:** accuracy figures use every guess made at the booth (1,092), including a few test games played by our team. Score figures come from the 71 players who left their results with us.
- **Sample sizes:** each clip was seen by between 1 and 21 players, so results for individual clips are not reliable on their own. The hardest-clips table only includes clips seen by at least seven players.
- **Content:** at the booth, real faces came from the FFHQ research dataset (the online version now uses licensed photos) and AI faces from StyleGAN. Real voices came from LibriVox and the Talrómur Icelandic speech corpus, and AI voices from ElevenLabs. Real video came from NASA, Pexels and Creative Commons vloggers, and AI video from Google Veo 3 and ByteDance Seedance. Full credits are on the [content credits page](/aiornot-credits).
## Play the game
AI or Not is now online at [sundra.io/aiornot](/aiornot). It works on phones and desktops.
---
# 5 Reasons Why Captioning Is a Powerful Marketing Tool
URL: https://www.sundra.io/articles/5-reasons-captioning-is-a-powerful-marketing-tool
Last updated: 2024-10-29
In the world of digital marketing, reaching a broad audience with your videos is essential — but it's harder than it sounds. Roughly 80% of viewers watch social media content without sound, 30–40,000 Icelanders live with some form of hearing impairment, and 18% of Iceland's population comes from an immigrant background. Add it all up, and captioning becomes one of the most powerful ways to increase the reach, impact, and shareability of your marketing content.
Here are five reasons why captioning your videos is a must for modern marketing.
## 1. Boost accessibility — and reach more viewers
Captions make your videos accessible to everyone, especially people with hearing impairments and those who speak Icelandic as a second language. With 18% of Icelanders from immigrant backgrounds and tens of thousands living with hearing loss, captioning can expand your potential audience by 20–30% in Iceland alone — and far more if you're publishing in multiple languages.
## 2. Capture attention in sound-off environments
More than 80% of people watch social media videos without sound. Captions grab attention instantly and let viewers understand your message whether they're on the bus, at the office, or relaxing at home. Without captions, most of your audience never hears a word you say — literally.
## 3. Drive more engagement and shares on social media
When viewers can grasp your message in the first few seconds — even on mute — they're far more likely to keep watching, interact, and share. Studies consistently show that captioned videos get higher engagement rates, longer watch times, and more shares, helping your content reach new audiences organically.
## 4. Improve your SEO and discoverability
Search engines can't watch videos, but they can read captions. Adding captions and transcripts gives platforms like Google and YouTube the text they need to index your content, surface it in search results, and recommend it to the right viewers. The result: more traffic, better rankings, and longer viewing sessions — all signals that further improve your visibility.
## 5. Build trust and stronger audience relationships
Captioning signals that your brand cares about every viewer, not just the ones who can hear or who speak your language fluently. That kind of thoughtfulness builds trust, strengthens your brand reputation, and deepens the connection between your company and its customers.
---
Captioning is one of the simplest, highest-ROI upgrades you can make to your marketing — it widens your reach, boosts engagement, improves SEO, and builds trust, all at once.
Ready to start captioning your content? [Book a demo](https://calendly.com/haukur-sundra/demo) or [get in touch with our team](mailto:hello@sundra.io) — we'll have your first video captioned in minutes.
