Audio to SRT subtitles
Turn a voice recording into a timestamped .srt for Premiere, DaVinci, CapCut and more, using your own Google Gemini key. Transcribe Sinhala or translate to English.
Files are processed in your browser and never uploaded.
How to use Audio to SRT subtitles
Add your file
Drop your audio, MP3, M4A or WAV file onto the upload area, or tap the area to choose it from your device.
Set the options
Choose Your gemini API key, Mode and Model. Every setting is on screen before you add a file, so you can see what the tool will do before it runs.
Run it
Press Generate SRT. The button stays disabled until there is something to work on, and each file shows its own progress and result.
What this tool does
Accepts audio, MP3, M4A, WAV and AAC files
Add an MP3, M4A, WAV, AAC, OGG or FLAC recording. Files under about 20 MB work best, which is roughly twenty minutes of speech.
Your own Gemini key, kept on your device
You paste your own Google Gemini API key and it is stored only in this browser. The audio is sent to Google's API under that key, so this is the one tool here that does use an outside service.
Sinhala subtitles, English subtitles, or both
Mode decides whether the recording comes back transcribed as Sinhala, translated to English, or as both files from a single run.
Choose the model that fits the job
Pick gemini-2.5-flash for speed or gemini-2.5-pro for accuracy, or type another model name. The result is a timestamped .srt for Premiere, DaVinci or CapCut.
Frequently asked questions
Is my audio uploaded anywhere?
Your audio is sent directly from your browser to Google's Gemini API using your own key, because that is what transcribes it. It does not pass through Don's Tools servers. If you would rather the audio not go to Google, this is not the tool for you.
How accurate are the timestamps?
Gemini's timestamps are approximate and can drift, especially on longer audio. The SRT lines up roughly, but expect to nudge some cues in your editor for exact sync. It is not frame-accurate.
Does it work well for Sinhala?
Gemini can transcribe Sinhala, but quality varies and is not guaranteed, especially for names, technical terms and fast speech. Always proofread the result.
Is there a file size limit?
The audio is sent inline and Gemini caps the whole request at about 20 MB, so this works best for clips up to roughly 15 to 25 minutes of MP3. For longer files, split them first.
Do I need to pay?
You use your own Gemini key, so any usage is billed to your Google account under Google's pricing and free tier. Don's Tools charges nothing and never sees your key.
Last updated: