Seply AI Speaker Separation
Seply is a browser-based AI audio tool built around speaker separation: it takes a single mixed recording of a conversation and produces an individual, time-aligned WAV track for each detected speaker. Rather than only labelling who spoke when, it renders each voice as a separate, editable audio file that can be muted, enhanced, re-ordered, or reused independently in an editing workflow. The workflow is upload, separate, preview, download. Users sign in, upload a recording of up to 1 GB, and let the system detect speakers automatically or specify the expected number themselves. Supported audio formats include WAV, MP3, M4A, AAC, and FLAC; supported video formats include MP4 and MOV carrying AAC or PCM audio, with compatible videos uploaded using only their original audio track without re-encoding. After processing, each speaker track can be previewed in the browser and exported as a WAV file, with all tracks starting at the same point so an edit stays in sync with the original mix. An Advanced mode is offered for recordings with overlapping voices. The homepage states 30+ languages are supported and that no installation or manual cutting is required. The site draws a clear distinction between speaker separation and speaker diarization: diarization tells you who spoke when and is typically paired with transcription, while separation creates an actual per-speaker audio file. Seply markets itself to podcasters, interviewers, meeting recorders, and other multi-speaker content producers, emphasising private processing, 30-day access to results, and credit-based pricing where the estimated cost is shown before processing begins. The tool is described as free to try, with users required to hold the rights to any audio they upload.
- Category
- Video
- Pricing
- Free
Problem solved
A single recording of a conversation mixes every participant into one track, so isolating, muting, or re-editing one person's voice normally means slow, error-prone manual cutting. Seply uses AI speaker separation to automatically detect speakers in a podcast, interview, meeting, or video and output a separate, time-aligned audio track for each one, so editors can work with individual voices directly instead of slicing the mix by hand.
Use cases
- Split a two-host or host-and-guest podcast into separate tracks for per-speaker editing and levelling
- Isolate interviewer and guest audio from an interview so answers can be trimmed or cleaned independently
- Extract individual participant tracks from a recorded meeting or panel discussion for notes and summaries
- Separate speakers from a video recording before mixing or re-balancing dialogue in a video editor
- Recover a usable single-speaker track when one participant's microphone was poor or noisy
- Produce per-speaker audio to accompany diarization or transcription output
Audience
- Content Creators
- Business Users
- Researchers
Pros
- No installation required; runs entirely in the browser
- Produces time-aligned WAV tracks that drop straight into an editing timeline
- Automatic speaker detection, with an option to set the expected speaker count manually
- Accepts both audio (WAV, MP3, M4A, AAC, FLAC) and video (MP4, MOV) inputs up to 1 GB
- Preview each detected speaker before deciding which tracks to download
Cons
- Requires signing in to upload and process a recording
- Free tier is limited; processing runs on a credit-based pricing model
- Upload size capped at 1 GB per file
- Output is limited to WAV files
- Overlapping speech can be difficult to separate cleanly and may still need Advanced mode plus manual review
Official website: https://seply.org/
Canonical LambdaBase page: https://www.lambdabase.com/tools/seply-ai-speaker-separation