Raw-first cloud recording

Recordings that fit your product, not just your call.

MediaSFU captures every participant separately, at their own high resolution, then composes the layout you choose. Record meetings, webinars, broadcasts, audio rooms and phone calls, control recording live, and keep the files in your own storage.

Recording00:12:48
Media
Video and audio
Audio from
All participants
Framing
Full display
Orientation
Landscape
Name tags
On
HLS package
On
Delivery
Your bucket
Example recording settings
4Output types: MP4, WAV, HLS and raw ZIP
7Webhook events from start to finish
$2Per 1,000 audio minutes, $6 at SD video
IncludedYour first composed layout
Layouts

Use our layouts, or record your own interface.

Most teams start with settings. When the recording has to match your product, your app can describe the layout itself.

Standard layouts

Pick your settings for the account, the room or a single recording, and MediaSFU composes the file.

  • Who is in itRecord audio from everyone, from people on screen, or from the host. Show all video or the main screen.
  • How it is framedFull display, best display, or both as separate files.
  • Which way upLandscape, portrait, or both.
  • How it looksName tags, background and tag colors, a line of text at the top, middle or bottom, and your watermark.
Recording settings reference

Custom UI recording

On request

Your recording follows the interface your users saw. Describe the main video, self-view, labels and badges once, then send small updates as the screen changes.

Example scene: guest camera, self-view and a status label
  • One scene format for web, mobile, native and game clients
  • Up to 3840×2160, 60 fps and 4 hours
  • Up to 16 video sources and 64 layers
  • A standard H.264/AAC MP4 at the end
  • A public example project with a scene validator and JavaScript and Flutter samples
Live control

Decide when it records, while the session runs.

In the room

The host starts, pauses, resumes and stops, and can change settings while paused.

From your server

Run the same actions over REST, so your own rules decide when a session records.

On phone calls

Pause, resume and stop recording for one call at a time.

In plain sight

Everyone in the room can see when recording is on, paused or stopped.

REST routes
POST /v1/meetings/:meetingID/recording/start
POST /v1/meetings/:meetingID/recording/pause
POST /v1/meetings/:meetingID/recording/resume
POST /v1/meetings/:meetingID/recording/stop

POST /v1/sipcall/:sipCallId/pause-recording
POST /v1/sipcall/:sipCallId/resume-recording
POST /v1/sipcall/:sipCallId/stop-recording
Outputs

Get the right file for the next job.

MP4

The finished video

Your chosen layout, ready to play, review or publish. Choose both orientations or framings and you get a file for each.

WAV

Audio on its own

For audio recordings such as podcasts, call review and speech processing.

HLS

A streaming package

Video recordings packaged as a ZIP of HLS files, ready for on-demand playback.

ZIP

The raw tracks

Each participant on their own, at their own high resolution. Turn this on for your own editing, AI or archive work.

Into your app

Know when it's ready, and get it where it belongs.

recording.startedrecording.pausedrecording.resumedrecording.stoppedrecording.processingrecording.completedrecording.failed

Webhooks

Follow each recording as it happens, with every event signed so your server can trust it.

Webhook guide

API

List recordings with GET /v1/recordings, filter by room name or sub-user, and set account defaults.

API reference

Dashboard

Set defaults, find finished files and download them, with nothing to build first.

Open recordings

Shared access

When you allow it, people can find a recording with its event ID and access code, without an account.

Find a recording
Storage

Keep recordings for as long as you need, in your own bucket.

MediaSFU keeps completed recordings for 72 hours. Connect your storage once, and files go there automatically.

AWSDigitalOceanWasabiCloudflare R2Custom S3
  • One bucketSend every recording to the account owner's bucket.
  • Per sub-userUse each sub-user's own bucket, and fall back to the owner's if it isn't valid.
  • What gets uploadedChoose finished files only, or raw tracks as well.
  • Who can open itDecide whether recordings and AI notes can be opened with an access code.
Pricing

Pay for what you record.

The first composed layout is included. Rates fall as more sources join, and stretches with nobody publishing aren't billed. A recording that captures nothing isn't charged.

  • First composed layoutOne orientation and one framing.Included
  • Both orientationsLandscape and portrait.+25%
  • All framingsTaken with both orientations, the total is +56%.+25%
  • HLS packageVideo recordings only.+5%
Open the recording calculator
Recording rates per 1,000 minutes
QualityResolutionRate
QHD2560×1440$18
FHD1920×1080$15
HD1280×720$12
SD640×360$6
QnHD320×180$3
AudioAudio only$2
Questions

Before you turn it on.

Meetings Webinars Broadcasts Audio rooms Record-only rooms Phone calls
What does raw-first mean?

MediaSFU captures each participant separately, at their own high resolution, instead of recording a flattened grid. Your finished layout is composed from those individual tracks, and you can keep the tracks themselves too.

How long do recordings stay on MediaSFU?

Plan on 72 hours. Completed recordings are deleted after that, so download them or send them to your own bucket if you need them longer.

What is the difference between standard layouts and custom UI recording?

Standard layouts are settings: choose who appears, the framing, the orientation and the styling, and MediaSFU composes the file. With custom UI recording, your app describes its own layout as a scene and updates it as the interface changes, so the recording follows your design.

How do I get custom UI recording?

It is enabled on request. Read the guide, look through the example project on GitHub, then contact us and tell us about your project.

Can I change settings once recording has started?

Yes, while it is paused. Pause it, adjust the settings and resume. The media type and HLS choice stay as they were when recording started.

How do I know when a file is ready?

Listen for the recording.completed webhook, then fetch the recording through the API or your bucket. The recording.failed event tells you when a recording didn't produce a file, including when nothing was captured.

Can each sub-user have their own bucket?

Yes. Recordings go to a sub-user's bucket when its settings are valid and fall back to the owner's. You can also send everything to the owner's bucket.

Do I have to keep raw files?

No. Delivering raw tracks is off by default. Turn it on when you want each participant's media for your own editing or processing.

Does recording affect encryption?

Media is end-to-end encrypted unless a backend feature such as recording is turned on, because MediaSFU needs the media to compose the file.

Where is recording not available?

Chat rooms, demo keys, and the Free and Connect plans. On meeting plans, each session can record for up to 1 hour on Basic, 3 hours on Standard and 10 hours on Advanced.

See what your recording should produce.

Try the settings, send a real request and check the file before you build around it.