Synthvids

Your Ai content growth engine

Design Tools

Calendar

Artificial Intelligence

The problem I wanted a way to build passive income through monetized, automated YouTube channels. Automating video content sounded simple until I actually tried to run it at scale. The bottleneck wasn't ideas or editing, it was rendering and time. I was paying $3.75 to $6 per short and $17 to $22 per long-form video using existing AI video tools. Even once a video finished rendering, there was still real time lost adding titles, tags and descriptions, and clicking through YouTube's upload flow to get it posted. I set out to build something for myself, but I built the pipeline with the right flags in place from the start so it could become a SaaS product one day if it turned out to be worth offering to other people. Why I built it I started by scraping content and using AI to turn it into YouTube videos, running everything myself through a chain of Python scripts. Somewhere along the way I realized I liked building the automation more than making the videos, so I kept building. I customized it for the problems I actually ran into as someone automating their own channel. When a clip at the end of a render came out wrong, I didn't want to re-render the whole thing, so I built a clip editor that shows every clip in a row of tiles, each one editable and regeneratable on its own. Scripts can be edited after generation, and hook styles (pattern-break hooks that stop scroll, delayed hooks, cold opens) are selectable right in your channel settings. I added the ability to prompt the script generator directly, since the script itself is what drives the visuals. Sound effects are generated based on what's actually happening in each scene. I wanted something fully customizable that also ran well on mobile, so I could post and scale channels while focusing on other things. It turned out to be genuinely fun to use, especially the ability to steer the AI visual generator through prompts. The cost problem got solved separately, and it changed the whole trajectory of the project: I found an open-source model that matched, and in my own tests beat, the quality I'd been paying per render for, running on hardware I controlled. That's what turned this from a tool I used for myself into something worth offering to other people. What's different here Synthvids handles short-form and long-form video. 768p is our base resolution across the board, and long-form also renders in native 1080p, the format YouTube favors for full-length video. 1080p costs the same as 768p in long-form, and is available on Starter plans and up. We hold that standard because it's what a long video needs to look right. When you make a video here, what we consider a video is a video, not a sequence of images with a voiceover and captions. You get a fully rendered video in whatever AI style you build. Advanced controls, custom hook styles, script prompting, per-clip regeneration, and render settings are there when you want them, hidden by default so the platform works just as well for someone who wants the simple path. We also offer a free short-from-long option: generate a long video and you get two finished renders in your Finished section, the long video and a short pulled from its best scenes, complete with meta, title, description tags and thumbnails, ready to post with one click or scheduled from your calendar. What's next The goal is to get people monetized and see what they build with it. I think the system is capable of better results than what I was getting when I built it, and I want to see what combinations of prompts and settings people land on, with a place to share what's working. We'll keep advancing the platform in phases after launch, including YouTube posting today, with TikTok and Reddit posting on the way as those integrations clear review. Who's behind it Synthvids is built and run independently out of California. That means fast iteration and direct support, not a ticket queue behind a big org. Thanks for being here early.