Start free

Webinar transcription.One recording, ten content assets.

Drop the webinar recording from Zoom, Teams, GoTo, or StreamYard. Get a speaker-labeled transcript, chapter markers, and an AI summary — ready to turn into blog posts, emails, and clips.

Drop a file, or pick one

MP3 · WAV · M4A · MP4 · MOV · MKV · OGG · OPUS · FLAC · WEBM — up to 100 MB anonymously

Paste a link, we’ll fetch the audio

YouTube · TikTok · Vimeo · Twitter · SoundCloud · Spotify · 50+ more

Record straight from your browser

Sign up takes 30 seconds — recording opens right after, in the dashboard.

No card required~90s per 60-min fileSRT · VTT · DOCX · TXTFiles auto-deleted in 24h

↓ Watch what comes out

60-minute webinar in. Repurposable transcript out.

Webinars are long, multi-speaker, and end with audience Q&A on mixed mic quality. We split host from panelist from attendee, mark chapters by topic shift, and surface the Q&A block as its own section.

StreamYard MP4 exportREC 1 host · 2 panelists · 62:14
auto-detected en-US48 kHz stereo · 192 kbps
~90s
Transcript · streaming94% accuracy · 4 chapters
S1

Welcome everyone — today we're walking through how mid-market teams are adopting retrieval-augmented generation in support.

S2

Thanks Priya. The thing we kept hitting was hallucination on policy questions, so we anchored every answer to a doc ID.

S3

Quick question from the chat — does that approach work for multilingual knowledge bases?

S2

Yes, with one caveat: you need the embeddings model to handle the target language natively.

94% on stereo webinar captureDOCX · SRT · TXT · JSON · chapters

↓ This is the dashboard

This is what loads when the job finishes.

Same layout as the real dashboard — Summary, full Transcript, Speakers tab, Exports. Key points and action items extracted automatically. Auto-tags on every job.

Try it on your own file — it's free

Three real options · honest comparison

Platform built-in. Otter or Fathom. Or us.

Zoom Webinar, GoTo, and Webex all ship a built-in transcript on paid tiers. Otter and Fathom join as a bot and dump notes into your CRM. We work with the MP4 you already have — including from StreamYard and Riverside, which don't ship a transcript at all.

Option 01

Webinar platform built-in

Inside Zoom Webinar, GoTo, or Webex. Locked to the host's paid plan and the platform's UI.

RequiresPaid webinar license
Speaker labelsHost only, often unlabeled
Languages1 per webinar
Chapter markersNone
AI summaryPaid add-on or upper tier
ExportVTT inside dashboard
Best forTeams that only need a basic caption file and never repurpose the recording outside the platform.
Option 02

Transcription.Solutions

Drop the MP4 from any webinar platform. Speaker labels, chapters, summary, and exports built for repurposing.

RequiresJust the recording file
Speaker labelsHost · panelist · Q&A
Languages99, auto-detected
Chapter markersTopic-shift detection
AI summaryTopics + Q&A, every plan
Cost · per min$0.03
Best forMarketing and content teams turning one webinar into a blog post, email sequence, social clips, and a knowledge-base entry.
Option 03

Otter / Fathom

A bot joins via calendar. Good for internal recaps, awkward for editorial repurposing.

RequiresCalendar OAuth + paid
Speaker labelsAcoustic, EN-tuned
LanguagesEN-strong, others weaker
Chapter markersAuto-headings (Otter)
AI summaryBehind paid tier
Cost~$17–20/user/mo
Best forSales-led webinars where the goal is CRM logging, not blog or knowledge-base content.

Pricing and feature flags reflect publicly listed plans as of 2026. Webinar platform AI summary availability varies by region and tier.

Specific to webinars

Four things that break generic transcription on a 60-minute webinar.

Webinars aren't meetings — they're long, scripted on one side, chaotic on the other. The defaults that work for a sales call don't.

What goes wrong

  1. 1Music intros and outros confuse the language detector. The first 30 seconds come back as nonsense, then the model resyncs late.
  2. 2Product names and speaker titles (CISO, RAG, SOC 2, your own brand) get spelled phonetically. Useless for SEO repurposing.
  3. 3Q&A block gets merged into the panelist's track because the attendee mic is quiet. You lose the question-answer structure.
  4. 4One long wall of text. No chapter markers, no way to find the moment a panelist said the quotable thing.

What to flip here

  1. 1Turn on Trim intro music in the job form. We skip non-speech audio at the head and tail before the recognizer runs.
  2. 2Paste speaker names, product names, and acronyms into Custom vocabulary. We pass them as recognizer hints, not hard matches.
  3. 3Set Speaker model: webinar (1 host + panelists + Q&A). The Q&A track gets its own diarization pass at lower SNR tolerance.
  4. 4Leave Chapter detection on. We mark topic shifts every 4–8 minutes — your editor jumps straight to the quotable section.

Recommended job settings for webinars

Drop a webinar MP4 and these flip on by default. Override per-job from the form.

Speaker model
Webinar · host + panelists + Q&A
Chapter detection
On · topic-shift based
Filler words
Removed (editorial mode)
Custom vocabulary
Speakers, products, acronyms
Summary
Topics + key quotes + Q&A list
Export
DOCX · SRT · TXT · chapters JSON

Accuracy · real-world numbers

94%+ on stereo webinar capture. Drops on phone-in Q&A.

The ceiling depends on how the webinar was captured. Studio-style panelist headsets recorded per-channel are the best case; audience questions dialed in over PSTN are the floor. Numbers below come from real customer webinar files, not synthetic tests.

96%+
Per-track studio capture

Riverside or Squadcast local tracks, each panelist on their own WAV. No bleed, no mixing artifacts — text-only error.

94%
StreamYard / Zoom cloud, host + 1-2 panelists

Standard webinar cloud recording at 128–192 kbps stereo. Diarization separates host from panelists reliably.

89%
Mixed Q&A over VoIP

Attendees unmuted to ask questions on browser audio. Mic quality varies; expect occasional misses on names and product terms.

84%
Phone dial-in Q&A

8 kHz narrow-band audio from attendees on the phone bridge. Usable for context, weakest on numbers, URLs, and acronyms.

Common questions

8 things people ask about webinar transcription.

01Can you handle a 90-minute webinar in one file?+
Yes. There's no hard duration cap — we've processed 4-hour conference recordings in single jobs. Turnaround scales roughly linearly, so a 90-minute file is usually ready in 6–8 minutes.
02Do you split the Q&A section from the main presentation?+
Yes, if chapter detection is on. We mark a chapter break when the conversational pattern shifts from monologue-with-interjections to question-answer. The Q&A block also ships as a separate section in the AI summary.
03Can I pull the recording from Zoom Webinar or GoTo directly?+
No — those URLs require a host login we can't impersonate. Download the MP4 from your platform's recordings page (30 seconds) and drop the file here. We don't ask for OAuth into your webinar platform.
04Will it work on StreamYard and Riverside files?+
Yes. StreamYard exports a single mixed MP4; we run acoustic diarization on it. Riverside exports per-participant tracks; if you upload them together as a zip, we use the per-track audio and skip diarization entirely — that's the 96%+ tier.
05How do I get a blog post out of this?+
The AI summary returns topics, key quotes attributed by speaker, and the Q&A list. Most customers paste that into their draft and pull verbatim quotes from the DOCX transcript. We don't auto-publish — the editorial pass is yours.
06What about attendee questions typed in chat, not spoken?+
Chat isn't in the audio file, so we can't transcribe what was never spoken. Export the chat log from your webinar platform separately and merge it manually — or send our live bot to the webinar, which captures chat in chronological order with the audio.
07Can you generate captions for the recording I'll re-upload to YouTube?+
Yes. Export as SRT or VTT — both burn-in cleanly on YouTube, Vimeo, and Wistia. The SRT respects speaker turns and chapter breaks, so caption timing matches the visual cuts.
08Is the recording stored long-term?+
No. The source MP4 is permanently deleted within 24 hours of job completion. The transcript, summary, and exports stay in your dashboard until you delete them. Webinar files are often gated content — we don't retain them.

Drop your webinar recording. Get ten content assets out.

30 free minutes every month. No card. Speaker labels, chapters, AI summary, all exports included.

Start free