MarketOtter is live - built in public
I've been posting here through the build, so here is the whole thing in one place. It's live today, and I'll answer anything in the comments.
WHAT IT IS
You tell MarketOtter what you're launching and who it's for. It does the research, works out the channels, helps you build the content and the schedule, and then holds you to that schedule. That last clause is the actual product — making a plan was never the hard part.
WHY I BUILT IT
Thirty years writing software for other people, with marketing always somebody else's job. Not a principled stance; I just found it uncomfortable, and uncomfortable work postpones easily when you have a backlog of things you're good at.
Then AI changed what one person could ship, and I shipped: an MIT-licensed PHP framework, a no-code builder on top of it, a writing platform, and this. Four products, and no idea how to get a single person to look at any of them.
THE DESIGN DECISION
Every AI marketing tool I tried was a blank box. A blank box quietly makes you the marketing expert — you still supply which channels, in what order, what to say on each, and when. Anyone who has that doesn't need the tool.
So MarketOtter predefines the structure: the channels, a credibility-first runway, the messaging arc, the schedule. AI is pointed only at the jobs it does reliably inside that frame. Less freedom for the model, better output. That trade held every time I tested it.
THE STACK
All of it built with Claude. ClaudePHP (MIT-licensed framework) → a no-code Builder on top of it → MarketOtter itself. Same principle throughout: predefine everything you can, so the parts that must be generated have somewhere solid to sit.
THE PROOF
This launch was planned and run by MarketOtter, on MarketOtter — every post in the runway, including this one. If it couldn't launch its own product, I'd have nothing worth telling you.
It's live at marketotter.com. 14-day trial, no card required.
Also live on Product Hunt today: https://www.producthunt.com/products/marketotter?launch=marketotter
AMA — the build, the stack, keeping AI-generated code maintainable across 57 modules, what didn't work, pricing, any of it. I'll be here all day.
Congrats on launching MarketOtter! It sounds like you’re addressing some critical pain points in the content creation and scheduling process.
When we were building our own solution, one of the biggest challenges was figuring out how to integrate research effectively with content creation and scheduling. We leaned heavily into automating the content generation while ensuring that it’s aligned with SEO best practices. For us, that meant building a system that not only produced content but could also evaluate it against performance benchmarks over time.
One thing I’d recommend is to gather user feedback early and often. We discovered that real-world usage unveiled challenges we hadn’t anticipated like the need for more intuitive scheduling features or better integration with existing tools developers were already using. It might also be useful to develop clear metrics users can track to measure the effectiveness of the content they're producing and scheduling with MarketOtter.
Also, consider how you could leverage AI to enhance the content creation aspect over time. We’ve found that when content is optimized for multiple channels, it dramatically increases visibility and engagement. If you can facilitate that kind of multi-modal optimization in your platform, it could really set you apart.
Best of luck and keep iterating based on user feedback!
Thanks for the comment. You are totally right about gathering user feedback early and often. Even dogfooding through it still doesn't catch many issues. Considering our findings so far, I think we should have spent more time with selected beta users before launching. A lesson for future launches.
It is interesting how often these kinds of experiences occur again and again, every team having to learn them. We read about them in all the business strategy guides and such, and then we push right on through and have to learn them again and again from personal experience.
It’s easy to understand the advice in theory, but launching makes the lesson much more concrete. Dogfooding can confirm that the product works for the person who built it, while beta users reveal where the assumptions, instructions, or workflow are unclear to someone seeing it for the first time.