I built JuzPost because I was the person it is for.

I have been running content properties since 2014. PremiumInfo came first, a tech blog covering software, cloud platform guides and HR tech, and CouponCruz followed in 2018 and grew into a high traffic deals site on AdSense and affiliate revenue. Along the way there were apps: EasieHR and Gameplux. Every one of them needed the same thing every day, which was the same post going out to every network. I did that by hand for years. It cost me about an hour a day, and the tools that could have fixed it wanted $50 to $150 a month before any of it had earned a dollar.

The other half of this is the day job. I joined CustomerLabs in June 2022 as an associate software developer and moved to data engineer a year later, working on a customer data platform in Chennai. One of the first things I shipped there was three integration systems that pushed user data out to TikTok, Google Analytics and Facebook, followed by a run of other API integrations with external systems. That is the same work JuzPost does, pointed in the other direction. I had been writing against these companies’ APIs for years before I wrote a scheduler on top of them.

The rest of that role was pipelines and scale. I rebuilt an event driven pipeline on Apache Beam Dataflow and took 50 percent off what it cost to run. I migrated 500 million user records from Firebase to Firestore and took another 20 percent off. I redesigned a BigQuery schema around materialized views for OLAP queries and cut query and storage cost by 30 percent. I built the logs pipeline on Dataflow and message brokers. None of that is glamorous, and all of it is the reason I am comfortable running a publishing queue that has to fire on time for other people.

What I have learned from both halves is that the platform APIs do not match their own documentation. Instagram’s app lets you post a carousel of 20 items while the API caps it at 10, so every scheduler on the market silently stops at 10. X counts a link as 23 characters no matter how long it is. YouTube measures a description in bytes rather than characters, so emoji cost four times what you think. Bluesky counts graphemes. Those are the kinds of facts that decide whether a post publishes or fails at the moment it matters, and they are what JuzPost checks before you schedule rather than after you have lost the upload.

So the rule I write by, in the product and on the blog, is that a limit gets checked against the live API or the published lexicon before it gets stated. When I cannot find a documented answer, the post says so instead of inventing one. That is the whole approach, and it is why I am building this in the open rather than describing it.

The short version of why JuzPost exists is on the homepage. You can check the working history on LinkedIn, what I am building on GitHub, and the day to day on X. Anything else, support@juzpost.com reaches me directly.