ShortsAIAn AI subtitle generator with word-level timing
ShortsAI generates subtitles with faster-whisper at word level, styles them for short-form video and burns them into the file with FFmpeg.
Captions can be translated: ElevenLabs dubs the audio into 30+ languages and the subtitles are regenerated in the target language so they match the new track.
Completely free, no watermark, no paid tier.
Why does word-level timing matter for captions?
Most free caption generators time subtitles per sentence, which is fine for a film and wrong for a short. Short-form captions are read one word or one phrase at a time, often with the sound off, and a caption that appears half a second before the word is spoken reads as broken. faster-whisper returns a timestamp for every word, and ShortsAI builds the caption blocks from those, so each block lands on the frame where its first word is said.
How do you add subtitles to a video automatically?
- Upload the video or paste a link. If you want clips as well, the clipper and the captions run in one pass.
- Processing transcribes the audio. Whisper handles accents, crosstalk and technical vocabulary well; a noisy recording is the main thing that degrades it.
- Choose a subtitle style in the editor: font, size, colour, position and the maximum characters per block.
- Export. The captions are burned into the video, which is what TikTok, Reels and Shorts need, since none of them reliably display an uploaded sidecar file on short-form video.
Can the AI caption generator translate subtitles?
Yes, and it does it the honest way. Translating a subtitle file while the audio stays in the original language gives a mismatch that viewers notice. ShortsAI instead dubs the audio with ElevenLabs into more than 30 languages, preserving the speaker's voice characteristics, then transcribes the dubbed track again so the burned-in subtitles are in the target language and in sync with what is actually being said.
Does it handle the safe zones on each platform?
Captions are positioned by default in the middle band of the 9:16 frame, above the area TikTok and Instagram cover with the caption text, username and action buttons. You can move them in the editor. The platform overlays differ slightly, so the middle band is the placement that survives all three.
What does it cost?
Nothing. ShortsAI is completely free with no paid tier, no watermark and no usage cap, at shortsai.app. You bring your own Google Gemini API key, whose free tier covers 1,500 requests a day; the site is supported by ads.
Subtitles are part of every clip; there is no separate charge and no per-minute caption quota anywhere. Dubbing is billed by ElevenLabs under your own key.
Common questions
- Is the subtitle generator free?
- Yes, completely: free with no watermark, no cap and no paid tier.
- Can I export an SRT file instead of burning captions in?
- The pipeline produces an SRT internally from the word-level transcript, and the API exposes it. Burned-in is the default because short-form platforms do not reliably show sidecar files.
- Which languages can it caption?
- Whisper transcribes most major languages. Dubbing and translated captions are available in 30+ languages through ElevenLabs.
- Can I change the caption style?
- Yes: font, size, colour, position and characters per block are editable per clip, with a live preview.