
Month 7, No Renewals, Shut It Down
Prigmar was my fourth product and the one that did not make it. Month 7, no renewals, I shut it down. Not a slow fade. A decision, written down, dated, on a Tuesday. That is how it ends when you are one operator with no funding and no team to blame. You look at the number and the calendar. Then you close the tab. The framework that came out of it is free today, and so is the kit.
The cost was concrete. Seven months of evenings. About $400 in tools. No office, no payroll, no investors waiting for a deck. Just my own hours, spent on a thing that did not earn. That is the part I still think about.
Nobody posts this part. The three products before Prigmar went fine. Tutoplus in 2012. BidLight. The internal automation system inside Tesla DGF. I had a run of things that worked, and I started to believe my own pattern. Prigmar corrected that in seven months, for the price of a used laptop and a lot of sleep.

The failure did not cost me the most. It taught me the most, and I only know that because I kept the evidence. If I had thrown away the repo and the notes and moved on, I would have paid $400 and seven months of evenings for nothing. Instead I paid that and got a framework.
If you run a church site or a small shop you made yourself, you already have a version of this file in your head. The mistake you keep making. The one that costs you a customer a year. Write it down tonight. It is worth more than the win you posted about.
What I Kept When Prigmar Closed
I kept two obvious assets: the customer list and the code. The third is the one that made the Free Builder framework possible. The decision log.
The decision log was plain. Every time I made a call that mattered, I wrote one line. Date, the call, what I expected, what it did. No prose. No reflection. Just the record. Pricing changes. Onboarding changes. The email sequence swap. Seven months of it, maybe 60 lines, sitting in a text file.
Read that file back in one sitting and the pattern is not subtle. The pricing mistake I made with Prigmar is Chapter 6 of the kit. I raised the price on a base that had not seen the product do anything yet. Two months of churn followed, and I blamed the market before I blamed the sequence. The log had the dates. The dates did not lie.

The onboarding email that lost me 40 users is a template in the 85-file MERN reference codebase. Not the good version. The failed version, next to the fix. Forty users hit that email flow and did not come back. I watched it in the numbers and could not explain it until I read my own notes from three weeks earlier.
That is the whole argument for keeping evidence. You do not need to get it right the first time. You need to keep the log. The operator who keeps the log and reads it beats the operator who makes the same call twice and forgets.
The Same Kit a Church Volunteer Needs, Minus the Guesswork
People ask me for a free starting point for church sites and classroom sites. Same answer every time. The tool is not the hard part. The decisions are. A church volunteer putting up a site on a Sunday afternoon has the problem I had in month 3 with Prigmar. No log. No way to know which call already failed.
The kit is free. The token is free. The course is free. You get the 206-page PDF, or the AI crawl token your AI worker reads to ship your SaaS. Same access either way. You pick the format that fits your week. A church admin with a laptop and two hours gets the same 12 module specs, at $0.
The 85-file MERN reference codebase is in there, including the failed onboarding template. The 12 module specs cover the pieces most people re-decide from scratch. The decisions you would otherwise make blind on a Sunday night, I already made and logged, including the ones I got wrong.

Someone running a church site does not need a monthly subscription to publish a page. They need to know which choices hold up and which ones quietly cost them the people who showed up once. That is a decision problem, and it is the one the kit solves.
Free means free. Pick the PDF or the token. Both cost $0. The only thing that costs money here is a $19 PDF about the next decade. Everything else is yours at $0, and you keep every customer and every dollar you earn.
Failures Are Assets When You Write Them Down
Working in public does not mean posting your wins with a screenshot. It means the failures are assets too, and only if you wrote them down while they were happening. A failure you remember is a story. A failure you logged is a product.
The stories that travel are rarely about the win. They are about the moment the person was wrong and said so, in specifics. Numbers, dates, what it cost, what they did next. Vague regret does not travel. A date and a dollar amount do.
That is also the honest answer to how to create a brand story that sells without feeling like a pitch. You do not invent one. You read your log and you pick the line with the most pain in it. For me that line is Prigmar, month 7, no renewals. I tell that story instead of the wins.

The same rule works for the brand topics your AI worker drafts. A church that admits the last campaign flopped and shows the corrected plan earns more trust than one that only posts the packed Sunday. Teacher sites work the same way. The class that did not land, and what you changed. People join the version of you that is willing to be specific about the miss.
The method is cheap. One file. Date, decision, expectation, result. Four fields. Do that for a month and you have more real material than most brands post in a year.
Keep the Evidence, Protect It Like Any Other Asset
A decision log is not just content. It is an asset, which means it needs the same care you would give a domain name or a customer list. Most solo operators think about protecting their records only after something goes wrong, and a log full of customer names and revenue lines is exactly the kind of file that matters.
Keep it private by default. The public version is the edited one, the lesson without the customer's name. You can tell the story of the onboarding email that lost 40 users without naming one of them. That is not a legal trick. It is basic respect, and it is also what makes the story worth telling.
Keep the raw file somewhere you control, not in a notes app tied to a service you might drop. A repo or a backup you can open in ten years. I still have the Prigmar log. That is the whole reason Chapter 6 exists, and it is the reason the failed onboarding template sits in the 85-file codebase next to its fix.

Treat the log as the source and everything else as output. Articles, emails, the pitch you give a new customer, the way you explain a price. All of it comes out of the same file. When the source is private and accurate, the output can be public and honest without exposing anyone.
Start it tonight. One line. If you already run a site for a church or a class, you have months of unwritten lines in your head. Get them out before you forget which ones were painful. The painful ones are the ones that sell.
You Direct AI Workers. You Keep Every Customer and Every Dollar.
Four products over 13 years and four industries: Tutoplus, BidLight, Tesla DGF, Prigmar. One operator, no team, no cofounder. Free Builder is the framework distilled from those four, and Prigmar is the one that shaped it most.
The model is simple. You direct AI workers. They are not a team you manage and pay. You point them at a task. You hold the decisions and the customers. The money stays with you. The kit gives them the reference material so they are not guessing, and it gives you the log so you are not repeating a call that already failed.
What you get at $0: the 206-page PDF or the AI crawl token, the 85-file MERN reference codebase, and the 12 module specs. The kit is free. The token is free. The course is free. If you want to read one paid thing, it is a $19 PDF about the next decade. That is the only price on the page, and I am not hiding it.

Free access is not behind a mystery form. The kit is free, you pick the PDF or the token, and you start. A church volunteer can have a working site plan before the coffee goes cold. A teacher can set up a class page without a subscription. A solo operator can point an AI worker at the specs and stop re-deciding what a contact form should do.
The takeaway is the one Prigmar taught me. You will not get it right the first time, and that is not the goal. Keep the evidence. Log the decision and keep the file. Then direct the worker and own the customer. The one that fails still pays you back. You kept the receipt. The kit is free. Go get it.
