Inky ClipsAdd captions to video automatically.
Inky Clips transcribes your video and draws every caption as it’s spoken, word by word, like a pen writing on the frame. It runs on your computer, so the video is never uploaded. Free to use.
The clip shows a scientist speaking, with captions written in Sans Bold as they talk.
Inky Clips listens and writes each word as it’s said. Pick a pen to restyle them.
Ink: White
How to add captions to a video
Open your video
Choose a video from your computer. It opens in the editor in your browser, and nothing is uploaded.
Click Auto captions
It's the speech bubble under the preview. The first time, the browser downloads the speech model, about 1.1 to 1.6 GB, and keeps it for next time.
Pick the spoken language
Choose the language people speak in the video. The one you used last waits at the top of the list.
Check and restyle
Click a caption on the timeline to fix its words. Long-press or right-click one and choose Edit all captions to change the look of every caption together.
Export
Export an MP4 with the captions drawn into the picture, or use Export subtitles to save an SRT, VTT or plain text file.
Word by word, in your font
Every word gets its own timestamp, so captions appear as they are spoken rather than a sentence at a time. They are grouped into short lines at pauses and sentence ends, which keeps them readable on a phone.
Each word writes itself onto the picture, stroke by stroke, in any of the 23 fonts: bold subtitle type such as TikTok Sans and Montserrat, or handwriting if you want the captions to look written. Set Writing to Synced to speech and the pen keeps pace with the voice.
Word-by-word captions, step by stepEdit them all at once
Click a caption on the timeline to correct its words. To restyle, long-press or right-click any caption and choose Edit all captions: the font, colour, size, outline, background and position you pick apply to every caption in the video.
Drag the captions on the preview to move them all together, for example above a TikTok’s buttons. Your own titles stay separate, so you can add a handwritten heading on top.
Add handwritten titles too68 languages drawn, 99 understood
The speech model understands 99 spoken languages. Drawing them is the harder part: every letter needs a pen path. Today captions draw in 68 languages, and we’d rather say so than show you boxes.
- 55
- Latin-script languages, accents included, in every font. English, Spanish, German, Polish, Latvian, Turkish and more.
- 13
- Russian, Ukrainian and other Cyrillic languages, Greek and Vietnamese, in TikTok Sans and Montserrat.
- 31
- Recognised but not drawn yet, such as Chinese, Japanese, Arabic and Hindi. You can still download their subtitles.
A subtitle generator, too
Want captions YouTube can switch on and off, or a file for another editor? After the captions are made, use Export subtitles. You get the same words and timings you corrected in the editor.
- SRT
.srt - For YouTube, Premiere Pro, DaVinci Resolve and CapCut. The most widely accepted subtitle file.
- VTT
.vtt - WebVTT, for web video players and the HTML video element.
- Text transcript
.txt - Just the words, for show notes, a blog post or a quick read-through.
Captions made on your computer, not ours.
Most caption tools upload your video and send the sound to a speech service. Inky Clips downloads the speech model to your browser once, then does the listening there. Your footage, its audio and its words never leave your machine.
Add captions to a videoQuestions
Is it really free?
Yes. Sign in with Google and caption as many videos as you like, with no minute limit. Videos you export on the free plan carry a small inkyclips.com in the corner; a paid plan from €7.99 a month removes it. SRT and VTT files are always free and unmarked.
Is my video uploaded to make the captions?
No. The Whisper speech model runs inside your browser. Your video, its sound and the captions stay on your computer. Our server only handles Google sign-in.
Which languages does it caption?
Whisper recognises 99 spoken languages. Captions can be drawn onto the video in 68 of them: Latin-script languages in all 23 fonts, and Cyrillic, Greek and Vietnamese in TikTok Sans and Montserrat. Chinese, Japanese, Korean, Arabic, Hindi and other scripts are recognised, but no font draws them yet.
How accurate are the captions?
Very good for clear speech in widely spoken languages. Music, crosstalk, heavy accents and smaller languages lead to more mistakes, and names are often misspelled. Read through the captions and fix any word before you export.
Can I get an SRT file?
Yes. Once the captions are made, Export subtitles saves them as SRT (for YouTube, Premiere and CapCut), VTT (for web players) or a plain text transcript. You can do that instead of exporting a video, or as well.
Can I change how the captions look?
Yes. Long-press or right-click any caption on the timeline and choose Edit all captions. Font, colour, size, outline, background and position then change for every caption at once.
Can it translate my captions?
No. Captions are written in the language that is spoken in the video.
Does it tell speakers apart?
No. Captions follow the speech in order, without speaker names.
Can I import an existing SRT file?
Not at the moment. Inky Clips makes captions from the speech itself.
Which browser should I use?
A current Chrome or Edge on a computer is fastest, because it can run the model on your graphics card with WebGPU. Safari works too, more slowly.