SUBTITLE GENERATION

How to generate subtitles from audio

Speech recognition uses the audio track, but the uploader accepts a video file. MP4, MOV, WebM, and MKV are the drop targets. A loose MP3, WAV, or M4A is rejected until you place that recording inside one of those containers.

You already have a finished picture? Upload that video directly. You do not need to strip the audio out. Generate subtitles from video

episode-04.mp4 is a valid upload because it has an audio track. notes.mp3 is not accepted until it is wrapped in a video file.

Do it in this order

  1. 01

    Put a bare recording into a video file

    If the source is only an MP3, WAV, or a voice memo, mux it into an MP4, MOV, WebM, or MKV. A still frame plus the audio is enough. Keep the same duration as the recording so the cue times match what you will publish.

  2. 02

    Upload that file, not the loose audio

    Drop the video container on the homepage or the Video to SRT page. The picture, even a still, stays in the browser. SubtitleGenerator extracts audio locally and sends that extract for transcription.

  3. 03

    Set the spoken language, then generate

    Choose the language on the recording. Podcasts with two languages need the language that carries the speech you want written down. Hinglish is the mixed Hindi-and-English option. There is no Auto mode.

  4. 04

    Review names, then export the file

    Fix Loop is where guest names and product words usually show up. Downloading the SRT requires Pro or Max. If you included a real picture and want captions burned in, export the video instead; Free can do that with a watermark at 720p.

What usually goes wrong

Dropping the MP3 and assuming it failed to transcribe

The file never enters the editor. The accept list is video/mp4, video/quicktime, video/webm, and .mkv. Wrap the audio, then generate.

Transcribing a stereo mix with music under the voice

Background music does not get its own track. Lyrics and speech can land in the same cues. Generate from the voice-forward mix when you have one.

Exporting SRT from a file that still has the slate

Cue times follow the file you uploaded. Trim the container to the episode before you generate, or every later cue will be late.

Questions this page should settle

Can I upload MP3 or WAV directly?

No. Put the audio in an MP4, MOV, WebM, or MKV, then upload that file. The transcription still runs on the audio track.

Does an audio-only container produce a captioned video?

You get timed cues either way. A burned-in export shows whatever picture is in the file. Use SRT when the destination is a podcast player or another editor.