Every edit is on screen — accept or revert with one click.
Drop audio or video file here
or click to choose · MP3 MP4 WAV M4A MOV
15 free minutes · no signup
25 languages · export SRT · VTT · TXT · DOCX · from $5
Upload once — pick what to download. Click a block to see an example.
click a block — I'll show you an example
File path: from recording to subtitles
First, exact verbatim transcription. Then AI cleans filler words and stumbles, fixes names and brands, restores punctuation — meaning is preserved.
Um, like, we were discussing with the open ai team you know the future of AI
afterWe were discussing the future of AI with the OpenAI team.
Lines aren't cut by character count. AI finds the phrase boundary — where the speaker pauses — within the language's specific CPL/CPS limits.
We'll discuss the new project with
the development team.
Preposition torn from its noun.
We'll discuss the new project
with the development team.
Break where the speaker actually pauses.
Each of 25 languages with its own CPL/CPS rules:
Speaker talks too fast, or the translation overruns the timing? AI compresses by meaning while preserving the substance. Don't need it — turn it off, keep the full text.
Actually, in principle, we could also consider that option here.
Viewer can't finish reading in time.
We could consider
that option too.
Fillers removed, meaning preserved.
Subtitles are ready — but the final word is yours. Every AI edit is on screen: accept or revert. A name fixed once is fixed everywhere it appears.
One upload — all assets together. Choose a format for each, optionally include timestamps and speakers. Take it as a single archive.
[00:12] Speaker 1
Tell us how the project idea came about.
[00:18] Speaker 2
We noticed teams were losing 2–3 hours a week to manual post-processing…
[00:34] Speaker 1
And how did you measure that?
topic
Q3 product strategy presentation to the board.
key ideas
figures
youtube-chapters.txtcopy ↗
00:00 Intro and introductions
03:45 Market overview and competitors
08:20 Q3 results in numbers
14:10 Strategy and next steps
19:50 Q&A
We need to talk about
the new project timeline.
Нам нужно поговорить о сроках
реализации нового проекта.
Russian runs ~35% longer. CPL blown, won't be read in time.
Обсудим сроки нового
проекта.
Reformatted for Russian limits. Timing preserved.
Prepaid balance in USD — no subscriptions, no hidden fees. The bigger the top-up, the bigger the bonus. Move the amount and watch it change.
Direct answers to what people usually ask before buying.
CPL — maximum characters per subtitle line. CPS — characters per second, i.e. reading speed. If a line is too long or text moves too fast, viewers can't finish reading. Limits vary by language — we keep them in one table:
| language | CPL | CPS |
|---|---|---|
| English | 42 | 17 |
| Russian | 39 | 14.5 |
| Spanish / French | 42 | 17 |
| German | 42 | 17 |
| Japanese | 23 | 3.4 |
| Chinese (simpl.) | 16 | 9 |
| Arabic | 40 | 20 |
CueAlign knows the limits for each of 25 supported languages and applies them automatically.
Three key rules: line length (CPL) within the language's limit, reading speed (CPS) within the language's limit, and semantically correct line breaks — at phrase boundaries, not mid-sentence. Maximum 2 lines per subtitle. The same foundational rules followed by professional standards like Netflix, BBC, and EBU-TT.
Standard transcribers give you raw text with filler words, stumbles, and name errors. CueAlign polishes the text, fixes names and brands, restores punctuation, breaks lines by meaning within language limits, and when needed fits the phrase to the timing. Every edit appears in a report — accept or revert with one click.
Translation is available between 25 languages. The key difference from standard translators: we don't just translate words, we reformat the result to fit the CPL and CPS limits of the target language. Russian runs 30% longer than English? AI compresses by meaning so the text fits the original timing. The source timing is preserved 1:1.
CueAlign reduces manual work significantly but doesn't eliminate it. Audio quality can be poor; AI can err on a rare name or specialized term. For such cases there's the change report — it shows all edits, names, brands, and numbers worth checking. Not "blindly trust the tool" but quickly scan the critical spots.
In a single upload: subtitles (SRT/VTT), polished transcript (TXT/DOCX), summary, YouTube chapters with timecodes, change report. The transcript optionally includes timestamps and speaker segmentation. Choose only what you need — or download everything as one archive.
Subtitles — SRT and VTT. Transcript, summary, and YouTube chapters — TXT or DOCX, your choice.
Files are processed in a secure cloud infrastructure and automatically deleted after processing completes. Audio and video are not stored permanently. Results are available for download for 24 hours. We never use your content to train AI.
Start with 15 free minutes. No credit card required.
Start subtitling →SRT · VTT · TXT · DOCX · 25 languages · from $5