The Sunday evening problem: why we built Clipr

The gap between intention and output

We started MRVL Technologies with a simple observation: churches and independent podcasters are drowning in content. A pastor delivers a sermon that speaks to dozens or hundreds of people. A podcast episode runs 90 minutes. A teaching video covers material that took hours to prepare. Then nothing happens. Not because the content isn't good, but because the friction between 'finished recording' and 'published short clip' is enormous.

The standard workflow went like this. Export the video. Import it into an editor. Watch the whole thing again (yes, really). Pick a moment. Mark in and out points. Hope the audio levels are right. Export it. Check the dimensions. Realize it's 16:9 and TikTok wants 9:16. Crop it or reframe it. Add captions. Export again. Then repeat four more times for different platforms. Most creators we spoke to simply skipped this. The content went nowhere.

When we dug deeper, the real problem wasn't laziness. It was that creators didn't have a team. A large media organisation can burn a producer to this work. A solo pastor, a podcast network with one marketing person, a church volunteer managing social media - they can't.

What changed when we watched the transcription happen

Early in building Clipr, we realised that the moment-finding part - the human listening to 45 minutes and knowing which 30 seconds matter most - was something we could help with. But only if we could turn the audio into text reliably, without sending it to the cloud.

We chose Apple Speech for on-device transcription. No cloud upload. No privacy concerns. No wait. The transcription happens quietly on your phone while you're doing something else. By the time you open Clipr, the text is already there.

Then we built a scoring system. It reads the transcript and flags moments that tend to perform well in short-form video: moments with emotional language, moments that stand alone without context, moments that answer a question or make a clear claim. The system isn't perfect. But it cuts the search time from 'listen to everything' down to 'review the top moments the system found.'

We tested this with a group of UK church social media managers in spring. One told us: 'I went from spending an hour finding clips to spending ten minutes reviewing them.' That felt like real progress.

Why we won't be the upload button

Some tools promise the whole chain: record, edit, post, done. We deliberately didn't build that. Here's why.

Every platform wants something slightly different from you. TikTok's algorithm favours a certain kind of pacing. Instagram Reels rewards captions styled a particular way. YouTube Shorts has different guidance again. If we automated the upload, we'd have to guess which platform you care about most. We'd almost certainly guess wrong.

So Clipr does the hard part. It transcribes on-device. It scores moments. It generates captions and bakes them into the video file. It reformats from whatever aspect ratio you recorded in into 9:16 vertical. On the Creator and Pro tiers, it removes the watermark so the clip feels like yours. Then it hands you a finished file and gets out of the way.

You spend two minutes uploading to TikTok with your own caption strategy. You upload again to Reels with their caption style. You upload once more to YouTube Shorts. That's three minutes of platform-specific thought, not an hour of that plus all the technical work. We found that creators are happy to do the upload step. They're happy to write a caption they believe in. What they hate is the whole video editing gauntlet before that.

The moment we realised batch processing mattered

In our first live test, a podcaster with a back catalogue emailed us. He'd recorded 20 episodes over the past two years and wanted to extract clips from all of them at once. Our system, at that point, handled one video at a time. Fair enough. He'd have to import, process, export, repeat. Twenty times.

He didn't bother. He processed two clips and left the rest.

That's when we built batch processing for the Pro tier. You can select up to five videos at once, and Clipr will work through them. The transcription still happens on your device. The scoring still ranks moments. But you're not babysitting each one. You can queue them up, leave your phone alone, and come back when they're done.

The same podcaster came back and processed his whole back catalogue in one evening. He posted clips for the next three months. Same content, same quality, completely different reach because it was actually being shared.

Free to start, features if you stay

We kept the free tier honest. Two clips a month with a watermark. Enough to try Clipr without committing. Enough to test whether it actually saves you time. If it does, Creator tier gives you 30 clips a month, removes the watermark, adds captions automatically, and the moment-scoring gets smarter. If you're processing back catalogues or managing clips for a church network with multiple channels, Pro tier unlocks unlimited clips, batch processing, and explanations of why each moment scored the way it did.

We've watched creators move through the tiers based on their actual need, not because we convinced them they needed more. That feels right.

What we learned about creators under time pressure

Building Clipr taught us that creators aren't lazy. They're busy. A pastor has a congregation to care for. A podcaster has the next episode to record. A church social media manager is also doing events management and volunteer coordination. They know that their recorded content could do more. They want to post clips. They just can't afford the time.

So we built for the constraint, not against it. Every feature in Clipr exists because a real creator told us: 'I would post more clips if I didn't have to spend [this much time] on [this specific step].' Transcription took time. Moment-finding took time. Captioning took time. Aspect-ratio changing took time. We removed each one.

The watermark was never about protecting our brand. It existed so free users felt the natural pull to upgrade. But every message we received from a free-tier user who'd extracted a beautiful moment and wanted to share it said the same thing: 'The watermark ruins it.' So Pro and Creator tiers ship without it. The clip feels like theirs. It should.