01
Capture
Preserve participant audio, video, screen share, and whiteboard sources instead of flattening the room too early.
Raw-first cloud recording
Cloud recording that stays useful after the call.
Preserve the source tracks, compose the view your audience needs, and deliver playback-ready or production-ready assets without rebuilding your recording pipeline.
- Keep isolated source media
- Compose multiple viewing formats
- Route output to your storage
Recording lifecycleOne capture, several outcomes
Video Audio Screen
Source tracks
Compose when neededLayout, labels, framing, watermark
Package & deliver
MP4WAVHLSRAW ZIP
Your archive or MediaSFU stagingAWS, DigitalOcean, Wasabi, or direct download
IncludedFirst composed layout
4Delivery families
72 hoursMediaSFU staging window
3Supported S3 providers
A recording pipeline, not a flattened file
Keep control from capture to archive.
MediaSFU separates source capture from composition, so the first playback file does not have to be the last useful version of the session.
02
Compose
Choose portrait or landscape, full-video or full-display framing, plus name tags, colors, text, and watermark placement.
03
Package
Create a ready-to-play MP4, audio WAV, adaptive HLS package, or raw-track ZIP for downstream production.
04
Deliver
Download from MediaSFU during the staging window or route assets into your AWS, DigitalOcean, or Wasabi bucket.
Choose the deliverable
Use the right output for the next job.
Deliver a finished playback asset now while retaining the option to edit, analyze, or stream the same session later.
Composed room video
A finished room view with your selected orientation and framing.
Best forPlayback, review, publishing
Audio-only export
A clean audio deliverable for rooms where video is unnecessary.
Best forPodcasts, QA, speech workflows
Adaptive streaming package
A packaged HLS archive with adaptive variants for streaming workflows.
Best forVOD and resilient playback
Raw participant tracks
Keep source media available for alternate edits and downstream processing.
Best forEditing, AI, compliance
Raw-first architecture
Recompose without asking users to record the room again.
Capture individual media sources once, then create the viewing format that fits review, publishing, adaptive playback, or post-production.
- First composition included: choose the primary portrait or landscape view.
- Alternate views stay possible: full-video and full-display variants preserve different framing priorities.
- Raw media remains useful: feed editing, transcription, AI analysis, or compliance workflows.
Two ways to operate
Start with the dashboard. Extend with code.
Operators can manage the workflow immediately; developers can automate the same controls as the product matures.
Configure in the dashboard
Set recording defaults, inspect completed files, download exports, and manage storage without building an internal control panel first.
Automate through API and SDKs
Define recording parameters with room creation, test payloads interactively, and keep recording policy aligned with your application lifecycle.
Storage ownership
Make retention a policy, not a surprise.
MediaSFU provides a 72-hour staging window for completed recordings. For longer retention, download promptly or send assets to your own supported S3-compatible provider.
Amazon AWSDigitalOceanWasabi
Fixed routingSend every recording to the account owner's bucket.
Variable routingUse a valid sub-user bucket, with owner-bucket fallback.
Upload policyChoose composed-only, raw plus composed, and HLS when enabled.
Access policyControl whether delivered objects are publicly accessible.
Workflow fit
Record the sessions that create lasting value.
Use recording where the output needs to be reviewed, replayed, published, or processed after the live room ends.
Meetings & conferences
Capture collaborative sessions and retain review-ready outputs.
Webinars
Record speaker-focused sessions with a consistent composed layout.
Broadcasts
Package live programming for replay or HLS delivery.
Audio-only rooms
Keep storage and processing focused when a WAV deliverable is enough.
Recording is not available for chat-only rooms.
Transparent expansion costs
Pay for the recording work you actually request.
Track capture is the base. The first composed view is included, then optional transformations are priced as small additions to track-recording cost.
Open the recording calculatorPrimary composed viewIncludedChoose a primary orientation and framing.
Additional layout variants+25%Of track-recording cost when extra compositions are requested.
HLS / VOD conversion+5%Of track-recording cost when adaptive delivery is enabled.
Before you configure
Cloud recording questions.
The operational details that usually affect architecture and rollout.
How long do recordings remain on MediaSFU?
Completed recordings are available in the MediaSFU staging area for 72 hours. Download them or configure an S3-compatible destination before that window closes.
Can each sub-user send recordings to a different bucket?
Yes. Variable routing can use a sub-user bucket when its configuration is valid and fall back to the owner bucket; fixed routing sends recordings to the owner bucket.
Do I have to keep raw files?
No. Raw upload policy can be configured for composed-only delivery, raw plus composed assets, and HLS when that conversion is enabled.
Which room types support recording?
Recording is designed for conferencing, webinar, broadcast, and audio-first workflows. Chat-only rooms do not support recording.
From test room to recording pipeline
See what your recording workflow should produce.
Configure the policy, send a real request, and compare the output before committing engineering time.