Drop a recording and get a transcript with speaker labels: every turn on its own line as [mm:ss] Speaker 1: what they said. Rename Speaker 1 to Julie once and it changes on every line. Export TXT, Word or SRT.
Drop the recording, or click to browse
MP3, MP4, MOV, WAV, M4A, WebM, AAC, OGG · up to 500MB
By uploading, you confirm you recorded this lawfully and have consent from everyone on it where the law requires. Terms
A speaker label in transcription marks who is talking at each change of voice. The transcript can tell voices apart, but it can't know names, so it numbers them in the order they first speak: Speaker 1, Speaker 2. Each time the voice changes, a new line starts with the time it began. You add the names.
Illustrative example with a fictional call.
All three are built in your browser from the transcript on screen, with the names you set. Copy all puts the TXT version on your clipboard.
| Format | What you get | Use it for |
|---|---|---|
| TXT | One line per turn: [mm:ss] Speaker: what they said | Pasting into notes, docs or a chat |
| DOCX (Word) | The same lines, with the time and speaker in bold | Sharing or editing in Word or Google Docs |
| SRT | Timed captions with the speaker's name on each | Subtitles in a video editor |
Labels come from how voices sound, so two people talking over each other, or two similar voices on one phone line, can end up under one label. Read the lines that matter before you quote anyone. And there's no bot that joins your calls: you need the recording file or a link to it.
A transcript is not a deliverable, though. Nobody forwards a transcript; they forward a slide with the number on it. If the transcript is a step toward a recap, uncheck the small talk on the topics step and send the interview synthesis, meeting recap or podcast show notes instead.
Yes. Transcribing, renaming speakers and exporting TXT, Word or SRT need no account and no card. Turning the transcript into a deck uses the free trial: you can preview the deck without signing up, and 2 decks are free.
Audio and video files: MP3, MP4, MOV, WAV, M4A, WebM, AAC and OGG, up to 500MB. Links work for Apple Podcasts episodes, Google Drive files shared with “Anyone with the link”, Dropbox, Loom download links and direct MP3/MP4 links. YouTube, Zoom, Vimeo and Spotify links are web pages, not files: download the recording and drop it instead.
Click the speaker's name on any line and type the new one. Every line by that speaker changes, and the exports use the new name. The names also carry over if you turn the transcript into a deck.
Some recordings come back without per-turn speaker labels. Then each block of the transcript is a topic, labelled with the person who spoke most in it, so a short reply from someone else can sit inside another speaker's block.
Yes. SRT is the plain subtitle format most video editors import. Each caption has its start and end time and starts with the speaker's name, using the names you set.
Uploaded recordings are automatically deleted about 24 hours after upload. The transcript text stays until you ask us to delete it at privacy@unfill.io. The AI and transcription services we use don't train their models on users' recordings, transcripts or decks. More on the security page.
Yes. Under the transcript, “Turn this into a deck, a follow-up email or social cards” opens the same recording at the topics step: uncheck what doesn't matter and you get an editable PowerPoint. Sales follow-up and Meeting recap decks also come with a follow-up email draft.
Try 2 decks free, no card required. Editable PowerPoint. No watermark.
Drop a recording