Video to Subtitles (SRT/VTT)
Upload a video, transcribe its speech with a free AI model, and download ready-to-use .srt and .vtt subtitle files.
Frequently asked questions
Are the subtitles actually time-synced to the video?
Usually yes - the speech model returns per-segment timestamps, which are used to build individually-timed caption lines. On the rare occasion a model doesn't return timing for a clip, the tool clearly says so and falls back to one untimed caption spanning the whole video, rather than guessing fake timestamps.
How is the audio extracted from my video?
Your video's audio track is extracted and converted to a small mono MP3 entirely in your browser (using the same ffmpeg.wasm engine as the Video to GIF tool) before it's sent for transcription - the video file itself is never uploaded.
Is there a length or size limit?
The extracted audio is capped at 15MB, which covers a good while of speech at the compressed rate used here, but very long videos can still hit that limit or take a while to process. Splitting a long video into shorter clips gets more reliable results.
Does it identify different speakers?
No, it produces a single transcript without speaker labels, similar to the Audio to Text tool.