Skip to main content
SIGNAL / PRIVATE RECORDING INFRASTRUCTURE

[ 001 ] -- For sessions that can't be redone

Never lose a second.

CraterSync records remote podcasts, audiobook narration and voiceover. Every voice is captured at full quality on the device it's spoken into and uploaded while you're still talking, so you get a clean, separate 32-bit track for each person. If the network dies mid-take, the take survives.

Pro $29/mo * 15 studio hours * guests join in-browser, no install

[ PICK A PLAN ] -- Same engine on every one

BUILT FORpodcasters/audiobook narrators/voice actors/voiceover studios
REC / 48kHz * 24/32-BIT WAV * UP TO 4K * VP9 / H.264
TRK_01 / CH_L+R / SEQ 00342
T+00:42:17.082
CHUNK00342 / 00509
UP94.2%
LOSS0.00%
SHAa8f3c91...

[ 002 ] -- One session, start to finish

One session, start to finish.

A host in Chicago. A guest in Lisbon on hotel Wi-Fi. An interview that only happens once.

SCENE / THE SESSION

The host sends a link. The guest clicks it -- no account, no install -- checks their mic and camera in the browser, and they're talking. Each voice records at 48kHz on its own machine, not from the call.

SCENE / THE DROP

Twenty minutes in, the hotel Wi-Fi dies. The recording doesn't notice. It never depended on the network: the pieces wait on the guest's laptop and upload the moment the connection returns. The call hiccups; the recording never does.

SCENE / THE EDIT

The session ends and each person is a separate track: 32-bit float, 24-bit WAV and FLAC, with the markers you dropped along the way. Nothing to sync, nothing mixed down -- open the stems and start cutting.

Narrating a book or directing a voice actor runs on the same engine: the script sits on the talent's screen and updates live, and takes are marked and circled as they land.

[ 003 ] -- The recording system

Six things every session does.

CraterSync is built around a single idea: the recording on disk is the product, not the feed in the browser. Everything else follows.

NODE 01

Local-first capture

Every participant records at source quality on their own device. Recording quality is decoupled from call quality -- your masters are never compressed to fit the network.

NODE 02

Progressive upload

Chunks travel to the cloud while you're still recording. By the time you stop, most of the session is already safely ingested. No multi-hour waits.

NODE 03

Separate tracks

Each voice on its own stem. Each camera on its own track. Edit with the same freedom a studio engineer expects -- no mixed-down surprises.

NODE 04

Session recovery

Network drops pause uploads, not recordings. Chunks are persisted locally and resume automatically the moment connectivity returns.

NODE 05

Per-session encryption

Every session gets its own 256-bit DEK. Chunks are sealed with AES-GCM before hitting storage. Retention ends with a cryptographic erase, not a support ticket.

NODE 06

One-link guest join

Send a link. Guests walk through a device check in their browser -- no account, no install. Hosts see permissions, bandwidth, and health at a glance.

[ 004 ] -- Built for the work that needs it

Pick the workflow closest to yours.

One recording engine. Each page below speaks to a specific shape of session.

Your recordings stay private

Nothing is published. There are no public download links and no feeds -- the stems live in your account until you take them.

Deletion is real

Each session is sealed with its own key. When retention ends, the key is destroyed and the files become unreadable -- not just hidden.

Nothing for guests to set up

Guests click a link and check their mic and camera in the browser. No account, no install, no password. You sign in with an emailed code.

[ EOF ] -- RUN A REAL SESSION

Record one episode. Judge the stems.

Get Pro, invite a guest, record, and download the separate tracks. Put them next to what you record with today and let your ears decide.