Five at once. The story behind Clipr's batch processing.

The spreadsheet problem

When we first released Clipr, most users were doing what that manager described: recording a sermon on Sunday, sitting down on Monday evening with one long video, and painstakingly processing it frame by frame. They'd hit export, wait, rename the clip, queue up the next one. A 60-minute sermon could become eight to twelve short clips, and the manual labour felt like punishment for being productive.

The feedback wasn't dramatic. It wasn't "your app is broken." It was quieter. People would use Clipr, extract three or four clips, then stop. Not because the clips weren't good; they were. But because the process exhausted them. They'd already given their week to ministry. The last thing they needed was a Monday night spent babysitting export dialogs.

We looked at our own spreadsheets. Users in the Creator tier were maxing out at 30 clips per month, but the actual pattern was stark: heavy use in bursts (Sunday evening, Monday morning), then nothing. The bottleneck wasn't the AI or the transcription. It was the human sitting in front of the phone, repeating the same six taps over and over.

The constraint that made sense

We could have gone bigger. Let users batch 20 videos, or 50. But we chose five, and that choice came from something specific. Most of the pastors and podcasters we spoke to didn't record 50 sermons a week. They recorded one, maybe two. A small church might batch their Sunday sermon with a midweek Bible study. A podcaster might process a week's worth of episodes at once. Five felt like the ceiling where the feature stayed useful without becoming maintenance.

More importantly, five videos gave us a technical floor. Batch processing introduces complexity: managing multiple transcriptions, scoring each one, queuing exports, handling the moments when one fails and four succeed. Five was large enough to solve the real pain ("I have three sermons to clip") but small enough that we could ensure reliability without it becoming a second app inside the first.

The decision also reflected who we're building for. Pastors aren't teams of editors with a new upload every hour. Church social media managers juggle a dozen tasks. A batch of five met them where they actually were.

What changed when we shipped it

The feature went live in the Pro tier in late autumn. Within three weeks, we noticed something in our usage data: the variance in clip production vanished. Users weren't clustering their output into frantic Monday evenings anymore. Instead, we saw more even distribution across the week. A Wednesday upload here, a Saturday one there. People were batching their recordings on their own schedule and letting Clipr work in the background.

Then came the messages. A church in Leeds told us they'd freed up five hours a month just from not sitting through the export-and-export-again cycle. A podcast producer in Belfast said she could now process a month's backlog in a single sitting, then move on to the thing she actually cared about: making the podcast itself better.

What surprised us most was how many people used it for archive work. Long-established pastors and teachers had years of recorded sermons gathering digital dust. Batch processing gave them a way to unlock that content without spending weeks in post-production.

The faith score explanation

We paired batch processing with something else in the Pro tier: faith score explanation. When the AI ranks a moment as engaging, you can now see why it chose that clip. The moment where a thought landed differently. The pause before an important sentence. The phrase that repeated three times because it was worth repeating. These aren't random numbers anymore; they're notes from the system about what made each clip distinct.

This matters because batch processing changes the relationship between you and the tool. When you're sitting there watching clips come out one at a time, you develop an instinct. You know which ones land. But when you're processing five videos at once and coming back to 40 clips later, you need that feedback. The AI becomes a collaborator, not just a processor. It explains itself.

What batch processing isn't

Worth saying plainly: this isn't a replacement for knowing your own content. Batch processing doesn't upload anything automatically. It doesn't post to TikTok or Reels while you sleep. You still take each clip, review it if you want, and decide whether it's worth sharing. What it does is remove the mechanical friction. The part where you're just shepherding video files through a series of steps instead of thinking about storytelling.

It also doesn't mean every clip is finished. Captions are baked in, the aspect ratio is fixed to 9:16, and the format is ready for social platforms. But the moment you export it, it's yours to use or refine. Some creators queue them all for upload. Others hand them to a team member with notes. Some watch through and delete two or three. The batch feature doesn't prescribe what you do next; it just makes the "next" part faster to reach.