Video Repurposing
Last week left a 52-minute kickoff call, a training, and a webinar fourteen people watched live. This agent works from copies of those recordings, reads the transcripts, and pulls the passages where somebody actually said something worth hearing — the payoff, not just the setup line. It cuts vertical clips with captions, writes the post copy, and lands the lot in your approval queue as drafts. It never posts.
Yours outright — no required subscription.
Every agent is built for your business — your systems, your approval chain, your way of working. We scope it on the call.
If this sounds too technical, don't worry. We take care of everything for you.
- Private VPS deployment. This agent runs on a server that belongs to you, not a shared cloud tenant. Your recordings are copied to your own box, and nobody else's footage sits next to them.
- A technician agent alongside it. Every hire ships as a pair: the Video Repurposing Agent that cuts the clips, and a VPS-technician agent that lives on the same box and keeps it healthy — patched, backed up, monitored. Video files are heavy, and somebody has to watch the disk. You're not hiring one thing, you're hiring a small team of two.
- Footage sources connected, off-limits ones excluded. Wired into wherever recordings already land, so nobody has to download and hand it files before a retention window eats them. At setup you name the folders it may never read — client calls, patient sessions, anything under an NDA or a consent that didn't cover marketing — and it has no route to them.
- Clips queued where you already approve from. Drafts land in the queue you already work out of, each one attached to its source file and its exact in and out timestamps. It hands off into your process instead of becoming another place to check.
- Security hardening. Private VPN, firewall, and encryption configured on your server before it copies a single recording.
- A live walkthrough of both agents. We sit down on a call and show you what the repurposing agent does, what the technician agent does, and exactly where the agent stops — so nothing is a black box on day one.
- 14 days of priority support after launch, to tune it against your real footage — your speakers, your supplier names, the way your calls actually sound. You own everything: the server, the agent, the code it runs on.
Not sure it fits? Check fit in 90 seconds in the free assessment chat.
Who this is for
- You have a folder with two years of recorded calls, trainings, and three webinars in it, and you've used maybe four minutes of any of it. Nobody has opened the March folder since March.
- You're the owner, you're the only person anyone wants to hear from on camera, and you will not schedule a shoot — but you'll talk for fifty minutes on a client call without noticing you're doing it.
- You have a marketing coordinator who burns a full day producing four clips, and that day is the first thing that disappears whenever a project goes sideways. Week three, nothing gets cut at all.
- You tried an auto-clipping tool and spent longer fixing clips that cut before the punchline and captions that misspelled your own company name than you'd have spent editing them yourself.
How it earns trust
The failure mode of AI in video isn't a bad cut — editors make those too. It's that a finished clip looks finished. Captions burned in, hook on top, and no way to tell what got left out. This agent makes the opposite bet: every clip arrives with the receipt that proves what it was cut from.
Every clip carries its source timestamps
The source file, the exact in and out points in the original, and the transcript passage it came from. So when a clip sounds off, you don't argue with it — you jump to 31:40 in the 61-minute recording and hear the whole thing in context before you approve.
The written caption sits beside the lines it came from
The post copy is shown next to the transcript it was drawn from. That's how you catch a caption asserting something the speaker never actually said: the two don't line up, and you can see that they don't.
A file log on your own server
Each clip records which source folder it came from, so confirming nothing was pulled from a folder you marked off-limits is a glance, not an investigation. And because the whole thing runs on your box, the log of every file it opened is a file you can read without asking anyone.
None of that replaces the last step, which is unavoidable by design: a person watches the actual clip, all the way through, before it reaches an audience.
Pairs well with
These share a workflow with this role. Tick any to add them to your setup.
We've got hundreds of hours of recorded calls and we've used about four minutes of it. Where does it start?
With the backlog, if you want it to. On the 15-minute call we pick which folders are in scope and which are off-limits, and it can work backwards through what's already there as well as forwards through last week's recordings. It works from copies — your original files are never modified, moved, or deleted.
I'm not getting in front of a camera. Does this need me to film anything?
No, and that's the whole point of the role. It only works from footage that already exists — calls, internal trainings, past webinars. If you never record anything, there's nothing here for it to do, and we'll tell you that on the call rather than build it anyway.
Half of what's on those calls I would never want public.
Then those folders get excluded at setup and it has no route to them. Two limits worth saying plainly: it reads audio and transcript far better than it reads what's on screen, so a job number on a whiteboard or an address in a shared screen can survive a clip that's clean of every name spoken. And it does not hold, request, or verify releases. If a customer, patient, or employee appears in a clip, clearing that is a human job that happens before approval, every time.
My marketing person burns a whole day to get four clips out. Can it just post them?
No. It never publishes — not to a channel, not to a page, not through a scheduler's post-now. Clips arrive as drafts and a person presses post. What it removes is the scrubbing, the trimming, the re-cropping and the caption cleanup; what stays is watching each clip through and saying yes. It also can't tell you the week's footage wasn't worth clipping — if nobody said anything quotable in fifty-two minutes, candidates still arrive and reviewing them still costs you time.
The webinar was good. Nobody has watched it since the day it aired. Can it use footage that old?
Yes — age doesn't stop it finding the good passages. But it cannot tell true-then from true-now. A webinar from March quoting a lead time, a price, or a headcount that changed in April will read as current the day it posts, because the clip carries no date a viewer can see. Every clip shows you its source file and timestamps for exactly this reason, and catching the stale ones is the approver's job.
What's covered under the $1,997 setup, and what isn't?
One company, one system of record for where clips get queued, one approval chain. Flat, one-time. Multiple brands with genuinely different approval rules, or a second queue with its own reviewers, is a scoped conversation on the call rather than a price adjustment. Typically live about seven days from kickoff.
Hiring more than one? A department on tap — the subscription puts a build team behind every request, agent after agent.