Conversation-aware transcription
Use speech to text to create an editable transcript that keeps who spoke and when. Seply handles overlap, tracks changes, and links each line to audio.
research-interview.mp4
42:18 · English · 2 speakers
What changed after the first prototype?
We stopped guessing and started listening to the calls.
Online transcription workspace
Use speech to text for audio to text or video to text files up to 1 GB or 10 hours. Online transcription adds language detection, word timing, labels, and private 30-day review.
MP3, WAV, M4A, MP4, MOV and more · up to 1 GB or 10 hours
Transcript settings
Label each detected voice in conversations and meetings.
1 Credit per started 3 minutes · minimum 1 Credit
Built for real conversation
Voice-to-text stops at words. Seply adds tools to review dialogue, correct attribution, verify audio, and prepare the result.
AI transcription can lose context during cross-talk. Seply keeps speech to text turns timed, linked to audio, and editable.
Noise affects results. Review linked audio before publishing.
Diarization groups timed words by participant to show who spoke when without claiming real identity.
Detection starts with neutral labels such as Voice A and Voice B.
What changed after the first prototype?
We stopped guessing and started listening to the calls.
2 detected voice
Focus on one participant, compare talk time, rename labels everywhere, or reassign a line. Corrections carry into copied text and exports.
Names and assignments remain editable.
13:48 · 56.9%
10:28 · 43.1%
Each passage keeps its timing. Click a timestamp to hear that moment, follow playback, or pause while searching a quote.
Word timing avoids full-file replay.
Current line · 18:43
We stopped guessing and started listening to the calls.
Correct wording beside the recording and save without changing tools. TXT, SRT, VTT, and JSON use the revised text and names.
Editable results stay private for 30 days.
Editable results stay private for 30 days.
Beyond a flat transcript
Basic speech to text returns paragraphs. Seply keeps audio, timing, attribution, and corrections connected after transcription ends.
Merged with little context
Timed turns stay in context
Generic paragraphs
Detected voices label each turn
Manual rewriting
Filter, rename, or reassign
Separate media player
Click text to hear the moment
Move into another editor
Edit and save beside the audio
Static text file
Revised TXT, SRT, VTT, or JSON
How it works
Choose supported audio or video. Large files upload to private storage, and the cost appears before processing.
Speech to text detects language, timestamps, audio events, and voice changes while keeping the recording available.
Search dialogue, filter voices, use the timeline, edit lines, and download the saved result in the format you need.
Built around spoken work
Separate participants, verify decisions against audio, and turn edited dialogue into dependable notes.
Keep hosts and guests readable, find quotes, and prepare named transcripts or subtitles.
Filter one voice, search answers, replay timestamps, and preserve wording for review.
Use speech to text for timed video captions, then correct and export SRT or WebVTT without rebuilding the timeline.
The terminology
In speech to text, speaker diarization labels timed segments by participant. It answers “who spoke when” without identifying the person.
Diarization labels text; it does not isolate tracks or verify identity. For separate audio files, use audio separation.
FAQ
Speech to text converts spoken audio into words. Seply keeps timing, labels, playback, editing, and exports connected.
Seply keeps detected overlap in context. Noise and recording quality affect results, so lines remain linked to audio.
It groups timed words by participant. Labels distinguish sounds in the file; they do not confirm identity.
Yes. Rename labels, reassign lines, add participants, or clear mistakes before saving.
Yes. Click a timestamp to hear that moment and follow the active line during playback.
Yes. Edit dialogue, save revisions, and use corrected wording in copied text and downloads.
Upload common audio and video formats. Download TXT, SRT, WebVTT, or JSON.
Source media and text remain private for 30 days or until you delete them.
Ready when the conversation ends
Upload, review timing, correct lines, and export text in one private workspace.