Audio to Text Converter

It is a tool that uses speech recognition to create a written transcript from a recording. ListenPad accepts common audio and video files, returns the text with optional timestamps, and exports TXT, DOCX, PDF, SRT, VTT, or CSV.

Drop a file into the upload area, pick a language hint or leave it on auto-detect, choose whether you need speaker labels, then start the transcription. The page shows the queue status and displays the transcript when the job finishes.

MP3, M4A, WAV, AAC, FLAC, OGG, WebM, MP4, and MOV are accepted. For a video file only the audio track is transcribed.

The default limit is 5 GB per file. It is a server setting, so a self-hosted deployment can raise or lower it, and the upload is rejected with a clear message when a file is over the limit.

By default a free account can submit a recording up to 30 minutes, and up to 3 hours on Plus and 10 hours on Pro or Max. Recordings that exceed the transcription provider limit are normalized and split into 10-minute chunks on the server, then stitched back into one transcript.

Yes. A free account includes 30 transcription minutes per 30 days and up to 2 uploads a day, stores your jobs in a workspace, and needs no payment details. An email address is all it takes to create one.

Yes. This server asks you to sign in before a recording can be uploaded, so every transcript belongs to an account rather than to a browser token. Registration is free with an email address, and the workspace then lists your recent transcripts with TXT, DOCX, PDF, SRT, VTT, and CSV downloads.

Yes. iPhone voice memos and other M4A recordings upload like any audio file. Pick a language hint if you know it, then download the finished transcript.

Accuracy depends on the transcription model configured on the server and on the recording itself. Clear speech and a single speaker close to the microphone produce the best result; heavy background noise, crosstalk, and strong accents lower it. Plan on reviewing the text before publishing it.

The upload form offers 34 language hints — English, Chinese, Spanish, French, German, Hindi, Arabic, Japanese, Korean, Portuguese, Russian and more, each written in its own language — plus auto-detect. The hint improves accuracy; the full language coverage comes from the transcription model the server is configured to use.

The speaker-label option returns separate speaker segments when the configured transcription service supports diarization. Results depend on audio quality, how much the voices overlap, and how many people are talking.

Yes. When the result includes timestamped segments, ListenPad formats them as an SRT or WebVTT download, or as a CSV with one row per segment if you would rather work in a spreadsheet. A transcript without timestamps can still be exported as TXT, DOCX, or PDF.

TXT is the plain transcript for writing and editing, while DOCX and PDF are laid-out documents with the file name, the date, and the AI summary when it is switched on. SRT is a subtitle file: numbered cues with start and end times, ready for a video editor or a captions upload, and VTT is the same thing in the format web players expect. CSV puts one segment per row with its start, end, speaker, and text.

The result panel is read-only so a finished transcript cannot be changed by accident. Copy the text or download TXT or DOCX, then edit it in your own document or editor.

Short recordings usually finish in well under a minute. Longer files wait in the queue, get split into chunks, and take proportionally longer, which is why the job runs in the background instead of holding the upload request open.

The job keeps running on the server. It belongs to your account rather than to the tab, so the workspace lists it with its current status whenever you return, and the finished transcript is waiting there with its TXT, DOCX, PDF, SRT, VTT, and CSV downloads.

The file is stored only while the job is processed and is deleted afterwards when the deployment is configured that way. The transcript itself stays in your workspace until the server’s retention window expires, which is unlimited for an account by default. Retention is a server setting, so publish the policy your own deployment actually enforces.

ListenPad does not use your recordings to train anything. The audio is sent to the transcription provider the server is configured with, so review that provider’s own retention and training terms before uploading confidential material.

Not from a URL. Download or export the audio first, then upload the file. Podcast episodes work well once you have the MP3 or M4A.

Yes. It runs in the mobile browser with the same upload flow, and there is no app to install. Recording files from a phone are usually M4A, which is supported.

New uploads are declined with a message telling you what is left. On a paid plan the allowance is tied to the period you paid for, so it resets in full when the next period starts rather than trickling back; on a free account it frees up as the rolling 30-day window moves. Upgrading on the pricing page raises the allowance as soon as the payment is confirmed. A job that fails never counts against your allowance.

Cancel from the workspace billing panel. Future charges stop and the plan keeps working until the end of the period you already paid for. If a renewal charge is ever late, access is held for a short grace period while the gateway retries before the account falls back to the free allowance.

Yes. It is a Next.js application with MySQL and a server-side transcription provider, deployed as a normal Node project. The upload limits, retention policy, queue concurrency, and provider are all environment settings.