Baian Caption
Turn recordings into translated subtitles.
Baian Caption uses context to translate recorded khutbahs, lectures and videos into SRT subtitles. Review and save the wording, then download each language.
The Caption app’s menus are in English and Arabic. You can choose other languages for your subtitles.How often Baian’s subtitles were preferredOver professional translations · English into each language
- Hausa 75%preferred
- Tamil 69%preferred
- Swahili 68%preferred
- Urdu 67%preferred
The line marks 50%: an even split.
How it works
From recording to subtitles in three steps.
- 01
Drop in your recordings
Add a whole batch at once. Each file can be up to 300 MB.
- 02
Pick your languages
Choose your subtitle languages. Every file in the batch gets them.
- 03
Download your subtitles
One SRT file per language, or all of them in a ZIP.
Inside the process: words, context and timing
Worked example
One clip.
35 languages.
Technical terms, big numbers and a joke, all in one long sentence. This passage from JFK’s Moon speech gives the subtitles something to work with.
- Upload your recording.
- Choose your languages.
Baian creates the subtitles. Review them, then download the SRT files.
Review note: the Chinese opening attaches the 240,000-mile distance to the rocket instead of the Moon. We’ve left the generated text visible.
About this example
Footage: NASA, 12 September 1962. John F. Kennedy speaking at Rice University, Houston. Excerpt: 15:22–16:34. Every language uses the same continuous passage.
Generated by Baian on 22 October 2026: an English transcript and 34 translations. We ran the languages in batches and regenerated Korean once after a review check. A Chinese rerun repeated the same distance error, so the original output is shown with a note. The generated wording is unchanged. Macedonian and Somali are in beta. This is one example, not an accuracy guarantee.
This public-domain excerpt is a subtitling example. It does not imply endorsement by NASA, Rice University or the Kennedy family.
After the download
How do you add them to your video?
Upload the subtitle file alongside your video in YouTube, or import it into your video editor. Each file contains the words and when to show them. Your original recording stays the same.
Add subtitles to YouTubeWhat you download
Welcome to today’s lecture.
- English.srt
- Arabic.srt
- Turkish.srt
Results
A closer look at Egyptian Arabic.
In a separate test, Baian was preferred over a professionally edited translation on 91 of 133 clips.
See all resultsAgainst a professionally edited translationEnglish to Egyptian Arabic · WMT26
- Egyptian Arabic 68%preferred
The line marks 50%: an even split.
Arabic transcription
Saudi Arabic mixed with English.
When you choose translation, Baian uses nearby lines to understand what the speaker means and express that meaning in your chosen language.
This test checks how accurately the spoken words were written down. Across 300 recordings mixing Saudi Arabic and English, Baian Caption’s word error rate was 9.5%. It does not score the finished translation. Results vary with the recording and dialect.
How this test was runSaudi Arabic mixed with EnglishTranscript words wrong · Perle test · lower is better
Questions
Questions about subtitles.
Where do I use Baian Caption?
Use caption.baian.ai to upload your recording. Baian is spelled B-A-I-A-N; baian.ai is our website and caption.baian.ai is the Caption app. Caption is for recorded audio and video. For an event happening now, use Baian’s live interpreter.
What files can I upload?
Upload video in MP4, MOV, WebM, AVI or MKV, or audio in MP3, M4A, WAV, OGG, Opus, AAC, FLAC or MKA. Each file can be up to 300 MB. Only upload recordings you own or have permission to process.
What do I get back?
One SRT file per language for each recording. An SRT is a small text file with each subtitle line and the time it appears. YouTube and most video editors accept it. Download the files one by one or together in a ZIP.
Can I upload several files at once?
Yes. Drop in a whole batch. Each file can be up to 300 MB, and every file gets subtitles in each language you choose.
Do I need to check the subtitles?
Yes. Check names, numbers and specialist terms before publishing. One regular customer told us they stopped checking every line because there was so little left to fix. That is their experience, not a guarantee for every recording.
Does it translate a live event?
This tool is for recordings. For an event that is happening now, use Baian’s live interpreter.
Can I edit the subtitles before downloading?
Yes. Open the completed job in Review, then switch to Edit to correct the wording. Select Save Changes before downloading one SRT per language or all files together in a ZIP. Wording edits keep each line’s start and end times. If you need to adjust the timing, do that in your video editor after downloading.
Does it handle Qur’an quotations in Arabic recordings?
For Arabic recordings, Baian can match Qur’an quotations to their verse references. It then uses the retrieved text when translating and reviewing the subtitles. Short or unclear fragments may be missed, so check the quotation and verse reference before publishing.
Your recording. More languages.
Upload a recording, choose your languages and download the subtitles.
Upload a recording The Caption app’s menus are in English and Arabic. You can choose other languages for your subtitles.Sources and methodsEvery number on this page, with its test and date
- Saudi Arabic mixed with English (Perle)Paper May 2026; Baian runs October 2026
Caption Tool 9.5% of words wrong, Baian’s live interpreter 15.6%, ElevenLabs 21.3%, OpenAI 43.7%, Azure 50.6%, Google 62.4%.
Perle, May 2026 (arXiv)Method
300 recordings of Saudi speakers switching between Arabic and English. Measures words wrong and meaning kept.
Perle’s authors tested ElevenLabs, OpenAI, Microsoft Azure and Google. We ran both Baian products on the same recordings.
Scored with Perle’s own rules and meaning model. The Caption Tool ran in transcription-only mode; Live used default settings with no language hint.
- Recorded speech recognition (Verli)Verli July 2026; Baian October 2026
1st in Spanish, Russian and Hindi; joint 1st in French and Japanese.
Verli speech-to-text accuracy, July 2026Method
200 FLEURS sentences per language. Measures words wrong.
Verli tested 15 or 16 engines per language. We ran Baian’s caption tool on the same test set.
Verli tested each engine as it was in July 2026 on its own 200-sentence FLEURS sample. We used the same test set and scoring rule on a different 200 sentences. Verli’s scoring rule splits Hindi words at vowel signs, which roughly halves everyone’s error rate, so we publish Hindi as a rank only.
- Baian against professional translators (FLoRes)October 2026
Caption Tool scored above the translators in Hausa, Swahili, Tamil and Urdu (preferred on 67% to 75% of sentences). Baian’s live interpreter had about the same overall quality in Arabic, Hindi and Korean.
FLoRes-200 professional translationsMethod
148 English sentences that also have professional human translations. Measures translation quality with Google’s MetricX-24 automatic judge.
Baian ran the test. The human translations come from the public FLoRes-200 set.
Baian worked from the spoken audio; the translators worked from text. We compared scores sentence by sentence and counted a result only when the 95% range did not cross zero and two runs agreed. “About the same” means the test found no clear difference; it does not prove the translations were identical.
- English speech to Egyptian Arabic (WMT26)15 October 2026
2nd of 30 systems, and preferred over the professional post-edited translation on 91 of 133 clips (68%).
WMT26 general translation testMethod
133 English YouTube clips, the speech test from the 2026 Conference on Machine Translation.
Baian scored every published system’s output, and its own, with Google’s MetricX-24 automatic judge.
Baian received the same automatic transcript as the other systems. WMT’s own human raters rank systems differently, so we always describe this as an automatic score.