Camera sync
Three cameras, three start times, one clock.
autoBlade reads the audio and lines them up. No clapperboard, no timecode box, no dragging.
Made for conversations. Built for Mac.
You bring the conversation. autoBlade syncs your cameras, transcribes every word, and makes the cut. Entirely on your Mac.
Free during beta · Apple silicon · No cloud required
Inside the app
Watch a real session in autoBlade, from importing the cameras to exporting the episode.
What it does
Sync the cameras, follow the speakers, and prepare the exports. autoBlade handles the repetitive work after a recording.
Three cameras, three start times, one clock.
autoBlade reads the audio and lines them up. No clapperboard, no timecode box, no dragging.
Speaker detection drives the switch, frame by frame.
When you both jump in it goes wide, then comes back. That is the four boring hours, gone.
A full transcript, timed to the edit.
Search it, quote it, ship it as subtitles.
Two hours of footage, turned around in about two minutes.
Built for the Neural Engine. No upload bar, no render queue you are waiting behind.
No cloud step, and no account needed to edit.
Unreleased interviews and NDA'd guests stay on the drive they were recorded to.
The same engine reframes to 9:16 and follows the speaker.
Burn in captions, drag, scale, restyle, then export the set.
The origin

several weeks back, I was spending several nights on an edit of a standard one-on-one podcast. A week later, I was staring down the exact same tedious project. I was so frustrated by the repetition...all I was really doing was seeing the audio level and cutting out the dead space. That should be a machine's job, not mine.
i then looked into existing software, but everything was clearly built by people who had never sat through the work.....(sadly)
so, we decided to build our own and called it "autoBlade". autoBlade works for any form of podcasting setup in existence all the way from Multicam to Shorts.
The app has a very simple UI with very few knobs to tweak, and does just one thing at its core! But at its best.
today, we have a fully working prototype. More than Prototype. I have fine tuned with 100s hrs worth of recording, internal testing and testing.
This will either end up big or crash and burn.
Beta offer · first 100 customers
Copy the code, paste it into the discount field at checkout, and Pro comes to nothing. It is good for the first 100 customers and expires with the beta, whichever lands first. Pro AI keeps the AI engine on every cut, and keeps its price.
USD · monthly or yearly
Pro
$13.99/ month
The editor itself. Sync your cameras and cut to whoever is talking, all on your Mac.
What’s included
Pro AIBest value
$29.99/ month
or $299 a yeartwo months free
Everything in Pro, uncapped, with autoBlade's AI engine on every cut.
What’s included
Monthly pricing, usage limits and where your footage is processed.
| Plan | autoBlade Pro AI | Descript Creator | Riverside Pro | Opus Clip Pro |
|---|---|---|---|---|
| Per month | $29.99 | $35 ($24 annual) | $29 ($24 annual) | $29 |
| Billing | Whole account | Per seat | Per seat | One price |
| Monthly volume | Unlimited | ~10 hrs per seat | Plan limits apply | 300 min per month |
| Your footage | Never leaves your Mac | Uploaded to cloud | Uploaded to cloud | Uploaded to cloud |
| Works offline | Yes, fully offline | Internet required | Internet required | Internet required |
Competitor list prices, September 2026. Check theirs before you decide.
Buying for an agency or team? Email losh@fromsilicon.com for volume pricing.
Apple silicon · macOS · free in beta
autoBlade is in beta and free to try. Hundreds of hours of real recordings have been through it already; you will still find rough edges, and I want to hear about them. 1.0 lands fall 2026, and every paid plan carries a 7-day money-back guarantee.
Questions
Still unsure about something? Email me directly. It reaches the person who wrote the app.
You give it your host, guest and wide camera files. It lines them up from the audio, transcribes the session, then cuts to whoever is speaking and goes wide when you talk over each other. You get a finished multicam edit plus a transcript, without touching a timeline.
A two-hour podcast takes under two minutes of your work: dropping the files in and picking your options. The system does the rest on its own while you get on with something else. Nothing is queued behind anyone else, because nothing leaves your computer.
No. Sync, transcription and speaker detection all run locally on your machine. The only thing that ever leaves is an optional text request to your own AI provider, if you paste in your own API key to unlock the shorts-finding feature.
Yes. The output is a real edit, not a locked black box, and you can take the render or the transcript into the editor you already use. autoBlade is there to kill the four hours of waveform-matching, not to have the last word on your episode.
Yes. The same speaker-detection engine reframes to 9:16, and you can burn in captions, drag them where you want them, scale the size and change the colours before exporting.
Finished enough that it edits real episodes end to end today; unfinished enough that you will find rough edges and I will want to hear about them. The full 1.0 lands in fall 2026, and beta testers get every update in between.