You can’t re-record a good conversation.
Your guest gave you an hour of their life and answered the question properly at minute forty. This is the studio built so that hour comes back at full quality, even if their Wi-Fi didn’t.
Free during the beta. Your guest needs a browser and a link.
The finished episode: the composed program plus a separate 1080p file for each person in the conversation.
A great hour that came back as a stuttering 480p mess
It is the same story every time. The conversation was the best you’ve recorded all year. The file is a smeared, freezing, out-of-sync version of it, because what got saved was the video call, and the video call is whatever the network could carry that afternoon.
Here the call and the recording are two different things. While you’re talking, your guest’s own camera and microphone are being captured in their browser at 1080p and written to their own machine, then uploaded in the background during the session. Their connection can wobble, drop, or die. The master doesn’t know about it.
- Crash and disconnect resume. Chunks are held locally and the upload picks up where it stopped
- Live upload health. You can see each person’s state the whole time: recording, uploading, safe
- Nobody babysits a progress bar. The take is usually safe before you’ve finished saying goodbye
The internet decides how the call sounds. It doesn’t get a vote on the recording.
Each participant showing a 1080p local master with its own upload state, including one that resumed after a dropped connection.
They agreed to an interview, not to installing software
The guest you most want is the one least likely to tolerate a download, an account, a plugin prompt or a calendar link that opens the wrong application. Every minute of setup is a minute of goodwill spent before you’ve asked the first question.
You send one link. They land in a green room, see themselves, pick the microphone they meant to use, and join. If they picked the wrong one, and they will, they can switch camera, mic or speakers mid-interview without leaving the room.
- Green room first. Device checks happen before anyone can see or hear them
- Reload-proof. If they refresh the tab, they land straight back in on the same devices
- Quiet host controls. Fix their display name, add a subtitle, unmirror their camera, all without asking them to touch anything
- A private chat. Send a link or a “two more minutes” cue that never reaches the recording
The green room: a camera preview with camera, microphone and speaker pickers, and a join button.
Booking to published, in one room
Nothing starts recording on its own. Every take is a deliberate press, and every take is kept.
- 1
Schedule it, or just start
Create a session as a recording and the live surface disappears entirely. No destinations, no stream keys, no way to accidentally go out to the world mid-anecdote.
- 2
Send the guest link
One link, reusable or regenerated. They do their device check in the green room while you finish your notes.
- 3
Press Record as many times as you need
Each start is a new take. Blow the intro three times and you keep all four; takes never overwrite each other.
- 4
Take the tracks
End the take and you get plain MP4s: the composed program, plus every participant’s own local master. Download them whenever you like.
- 5
Cut the versions
Transcript, text-based edit, chapters, an audio mixdown for the feed and vertical clips for the algorithm, all from the same recording, without moving it to a fourth tool.
Separate tracks, because someone always coughs
A single flattened video is fine right up to the moment your guest’s dog barks over your best question. Because every person was recorded separately, you can duck one voice, re-frame one camera, or lift a whole answer out, without touching anyone else.
The transcript comes labelled by speaker, because we already know who was on which track. Delete a sentence in the text and the sentence goes from the video. Cut the “ums”, drop the section that didn’t work, reorder two takes into one episode.
- One file per person at 1080p, plus the composed program
- Speaker-labelled transcripts. No guessing who said what in a mixdown
- Editing by deleting text. Cut filler and whole segments from the transcript
- Plain MP4s, no expiry. Take them to whatever editor you already use, whenever you want
A speaker-labelled transcript with one line struck through, showing a cut made by deleting text.
The interview is one recording. The show is five things.
Video podcasting isn’t a second workflow bolted onto an audio one. It’s the same hour, exported in the shapes each platform wants.
The video episode
The composed program with your layout, branding and lower-thirds already burned in, ready for YouTube without a round trip through an editor.
The audio episode
An audio-only mixdown for the feed, off the same take, so the podcast and the video are never two separate recording sessions.
Clips for the feed
Highlights picked from the transcript, cropped to 9:16, 1:1 or 16:9 with captions burned in. That is the part of the job that usually eats a whole afternoon.
Notes and chapters
Show notes, chapter markers and title options drafted from what was actually said, not from what you can remember at 11pm.
A show that looks like a show
Seven layouts switched live, five overlay types at nine positions, a ticker, your background, accent colour and logo. Not a video call with two rectangles.
A live episode, when you want one
The same room streams to YouTube, any RTMP host and your own site. Record it at the same time by ticking the box.
Your guest will not sit for a second attempt.
Every other failure in this job is recoverable. A bad thumbnail can be redrawn, a bad title rewritten, a bad edit re-cut. A lost interview is just gone, and asking a busy person to do it again costs you the relationship as well as the hour.
What podcasters ask first
What does my guest actually have to do?
Open a link in their browser on a laptop and allow the camera and microphone. No account, no download, no plugin, no app to approve. They get a green room to check their devices before anyone sees them.
What happens when their internet drops mid-answer?
Their local recording carries on, because it never depended on the connection. They reconnect automatically, and their upload resumes from the chunk it stopped on. You’ll see it go from recording to uploading to safe.
What if they close the laptop before the upload finishes?
The captured chunks live in their browser’s storage, not just in memory. Reopening the session page resumes the upload. You can see, live, whose files are still in flight.
Can I get an audio-only version for the podcast feed?
Yes. Export packs include an audio mixdown and the per-person track files, so the audio show and the video show come out of one session rather than two recordings and two calendars.
Do I have to publish the video, or can I record video and only use the audio?
Entirely up to you. Recording video costs you nothing extra here, and it means the day you decide to put the show on YouTube, your back catalogue already exists.
Do recordings get deleted if I stop paying, or after a few weeks?
No. There’s no delete window and no meter on recording hours, takes or downloads. The files are plain MP4s you can pull down whenever. See recording for how the takes are stored.
Where to go next
Book the guest. We’ll take care of the hour.
We’re opening rooms a handful at a time, free while the beta runs. Leave your email and we’ll open one for your next interview.