Pinned Tweet
1/2 Still in progress, but yesterday I announced bisibility billing - a $3/month subscription for app access starting next month. Access will be free through the end of the year for all my X followers - just DM me.
If you had $500/month in Claude/ChatGPT API credits and had to spend it all on SEO, what would you build or automate first?
bisibility is live on Product Hunt today! 🚀
producthunt.com/products/bis…
Open-source SEO for rank tracking, keyword research, and backlink analysis.
I’d love your feedback in the comments on the launch page 🙂
🤖 Made with AI
I’ve just added a new feature to the bisibility dashboard: track your Domain Rating (DR) for free with the Ahrefs integration.
i've heard a lot of saas founders share fake stats. so here are some real ones to keep the balance.
all time: 87 users, 80 projects, 1,013 keywords
last 30 days: 52 users (+271%), 49 projects (+250%), 215 keywords (+6%). ~10 accounts active in the last 7 days.
current mrr: $3 (my test payment; should be excluded ;)
screenshot from my internal dashboard
I am preparing to launch on Product Hunt, and I find SubmitMap.com useful for finding other places to post the SaaS. It also has clear instructions on submitting a site and preparing the data most sites need. I highly recommend it to anyone who wants to build strong backlinks and get more users.
I’ve been waiting two weeks for Apple to even start reviewing my iOS app (a cute mobile game I’m making just for fun). That’s ridiculous - lots of people on Reddit are seeing the same delays. Official guidance says it should usually start in about two days, which is why I chose the App Store over Google Play. Right now it looks like there are too few reviewers and way too much AI slop in the queue.
4 seo trends i keep seeing on x this week.
1. people build their own stack: claude code + dataforseo + search console. way cheaper than ahrefs or semrush, and you control every query.
4. agents make up numbers when the api doesn't flag missing data. averaged ratios, partial days, rows that never existed.
the diy stack works. the catch: results live in a chat session, close it and the history is gone.
worth having a tested dashboard next to it that keeps every run. that's what i'm building with bisibility: bisibility.com
You can now limit your AI agents’ requests to paid APIs using usagekit.dev, so you won’t wake up to a huge bill anymore.
It was extracted from Bisibility, and you can use it in at least two ways:
- Build your own API wrapper product with a bring-your-own-key (BYOK) option.
- Run a dashboard and ask your agents to route their requests through it.
It’s still a work in progress, but you can already join as an early tester. Enjoy!
Bisibility just reached 100 stars on GitHub. Still a lot of work to do, but it’s nice to see people interested in the project. Thanks to everyone who starred it!




