My wife survived domestic violence. Then her ex found a new weapon.

She thought leaving meant it was over

Leaving an abusive relationship is not a single moment. It's a process that takes enormous courage, planning, and sacrifice. My wife did it. She got herself and her two children out. She rebuilt. She moved forward.

Most people who haven't been through it assume that once you leave, it's over. You're safe. The nightmare ends.

What they don't tell you is that for many survivors — especially those with children — leaving is just the beginning of a different kind of fight. Because for a certain kind of person, losing control over someone they believe they own is completely unacceptable. And if they can no longer control you with fear, they will find another way.

Her ex found another way. He discovered that the family court system — with its mountains of paperwork, its overwhelmed judges, its expensive attorneys, and its endless procedural loops — could be turned into a machine that ran indefinitely. A machine that could keep her exhausted, keep her broke, keep her uncertain, and most devastatingly: keep her separated from her children.

He didn't need to hurt her anymore. The court process did it for him — filing by filing, hearing by hearing, year by year.

There's a name for this in the research literature. Litigation abuse. Post-separation coercive control. The deliberate use of legal proceedings not to resolve a dispute — but to maintain dominance over someone who had the audacity to leave.

We didn't know the name for it when we were living it. We just knew it felt like drowning in slow motion.

The filing never stopped

By the time I was deep in it alongside her, there were hundreds of court documents spread across multiple cases in multiple jurisdictions. Motions. Counter-motions. Emergency applications. Protective order requests. Status reports that read more like accusations. Affidavits filled with claims that had never been tested, never been compared, never been put side by side with anything else in the record.

Our attorneys were good. Several of them. They worked hard and they fought hard. And one by one — worn down by the volume, the cost, the relentlessness of it — they stepped back. By the time I was fully in it alongside her, my wife had spent over $200,000 in legal fees. Two hundred thousand dollars. Not to win anything. Just to keep showing up. Just to not lose by default.

Family law is expensive under normal circumstances. When the other side treats litigation as a lifestyle, the financial attrition is part of the strategy. Bleed them dry. Make every hearing cost something. Make every motion require a response. Eventually, most people run out of money before they run out of fight — and he knew it.

Meanwhile, my wife was watching the space between herself and her children slowly, deliberately, methodically shrink. Visits were made difficult. Then supervised. Then supervisors were driven away. Then her absence from visits — absence caused by barriers he had constructed — was cited as evidence that she didn't care about her children.

It was a trap. And it was working.

Every new filing was another brick in a wall designed to separate a mother from her children. And because each brick arrived separately, no one had ever seen the whole wall.

I'm not a lawyer. I'm someone who reads data.

Twenty years in tech and digital marketing. In recent years, AI became central to how I work — building systems that find patterns in large, messy, unstructured information. I've spent my career doing exactly one thing: taking data that looks like chaos and finding the signal hidden inside it.

One night, sitting with a folder of hundreds of court documents, I had a thought that changed everything.

I had been reading these filings the way everyone reads them — one at a time, as individual documents. But that's not what they are. They're a record. A log. A dataset. And the most important thing about a dataset isn't any single entry. It's the relationships between entries. The contradictions. The reversals. The moment where what someone swore was true on page 23 of one filing completely falls apart when you put it next to page 47 of another.

A human attorney cannot hold 400 documents in their head simultaneously. They don't have to be incompetent for contradictions to slip past them. They just have to be human.

I stopped reading the filings as paper. I started reading them as data. And then I built a system that could read all of them at once.

What we built

I started with a single filing. Asked an AI to extract every factual claim in it. Then did the same with another. Then asked the AI to compare them.

The first pass took about an hour. It found three things no attorney had ever raised — not because our attorneys were careless, but because no one had ever read those two specific documents at the same time, side by side, looking for conflict.

I kept building. Over the following weeks I created a full system — an AI-powered vault with every filing ingested, indexed, and cross-referenced. Query the entire case history in plain English. Ask "what has he said about the children's refusal to visit?" and get every relevant passage from every filing, with exact document names and page numbers.

What I built in those weeks gave us something we had never had in years of litigation: a complete picture. Not fragments of it. Not what one attorney remembered from six months ago. The whole thing, searchable, comparable, and mapped.

But the part that changed everything for us in court wasn't just finding the contradictions. It was what we could do with them immediately after.

The system doesn't just surface a lie — it drafts the motion around it. You ask it to write a response to a filing, and it produces a professionally structured document: proper legal formatting, every argument supported by citations pulled directly from the case record, exact page numbers, exact document names, exactly the way a court expects to receive them. What used to take an attorney days of billable hours — reading, cross-referencing, drafting, revising — the system produces in minutes.

For our attorneys, it meant they could focus on strategy instead of archaeology. For the periods when we were navigating parts of the case ourselves, it meant we could file responses that looked and read like they came from someone who had read every single document in the file — because the system had.

What the AI found

I want to share three of the patterns the system uncovered. No names, no identifying details — because what happened in our case is not unique. These patterns show up across family courts in every state, every day, in cases where one party has decided that truth is optional and volume is strategy.

He had spent years building a case file designed to make my wife look like the problem. The AI read all of it and showed us exactly where it fell apart.

Then he found out she got married

Through everything — the years of filings, the drained savings, the supervised visits, the attorneys coming and going — my wife had never stopped moving forward. She had rebuilt her life. And when we got married, something in him snapped.

The harassment that had never fully stopped suddenly shifted into something different. More targeted. More frantic. The volume of filings increased. The allegations became more extreme. And for the first time, he turned his attention away from my wife and directly toward me.

I think, in his mind, the marriage was the final proof that he had lost control. She had moved on. She had someone in her corner. She wasn't alone and exhausted anymore. And that was intolerable to him.

So he came after me. A protection order application claiming I had done things I had never done, supported by claims that had no basis in fact. Then a lawsuit. Then a criminal complaint.

My first instinct was the same as everyone else's: hire an attorney. I did. A criminal attorney, because that's what the situation required. And the bills started coming in immediately — thousands of dollars, on top of everything my wife had already spent, on top of everything we were already carrying.

But something else was happening at the same time. The more I read his filings — really read them, the way I had learned to read them — the clearer a picture emerged. This wasn't just someone who was angry or vindictive. There was a pattern to everything he did that went deeper than strategy. The way he shifted positions the moment they stopped serving him. The way he weaponized every process meant to protect children. The grandiosity. The complete inability to accept that he had lost control. The absolute certainty that the rules applied to everyone but him.

I'm not a psychiatrist. But I had read enough — in the documents, in the research on high-conflict litigation, in the forensic evaluators' own filings — to recognize what I was looking at. And once I saw it clearly, something shifted in how I approached everything.

I stopped being afraid of him. And I started working.

I spent weeks building out the AI system — mapping every claim he had ever made against me across every document in the file. Every allegation, cross-referenced against prior sworn statements. Every "fact" he had presented, checked against the actual records. The contradictions stacked up fast. Not because he was careless. Because when you lie habitually and strategically across hundreds of documents over years, the lies eventually contradict each other. You can't keep the story straight forever. The AI doesn't forget. It doesn't get tired. It just keeps reading.

And then I made a decision that surprised even me.

I fired my attorney. I decided to represent myself.

Not because I thought I was smarter than a lawyer. But because I had something no attorney in this case had ever had: I had read every single document. I had mapped every contradiction. I knew exactly where his story fell apart, and I knew exactly how to show it. I drafted a detailed motion laying out the complete pattern — with citations, with page references, with his own words set against each other in sequence. Then I walked into that courtroom and I cross-examined him myself.

I already knew what he was going to say. I had read everything he had ever said. And when his answers contradicted the record — which they did, repeatedly — I had the documents in my hand to show it.

The protection order was denied. The judge found his allegations not credible and specifically noted a motivation to lie. The lawsuit was struck down — and when he appealed, the appellate court affirmed. The criminal complaint went nowhere.

I won. Not with a legal team. With a system, a thorough preparation, and the complete record of everything he had ever put in writing.

I walked into that courtroom knowing exactly what he was going to say — because I had read every word he had ever filed. He didn't stand a chance against his own record.

Then the courts started seeing what we saw

When independent, court-appointed professionals finally had the time and access to do their jobs without being immediately attacked, they found the same things our AI had already mapped.

One evaluator described what she observed using words directly from the forensic literature on coercive control. Another noted children whose behavior in visits looked coached — not distressed, not genuinely reluctant, but performing a refusal that had been rehearsed. A third identified restrictive gatekeeping: barriers to visitation constructed and then cited as evidence of abandonment.

The AI had flagged all of it, months before any professional touched the file.

Why I built CaseBrake

At some point during all of this, I started thinking about all the other families going through the same thing without what we had.

Because we were fortunate in a strange way. I had twenty years in tech and digital marketing — and in the last few of those years, AI had become the core of how I work. I knew which tools to use. I knew how to build the system. And we had already paid $200,000 in legal fees — which meant we had learned, the expensive way, exactly what the system can and cannot do for you on its own.

Most people in this situation don't have any of that. They're not tech professionals. They haven't spent $200k learning the hard way. They're just parents — terrified, exhausted, watching their savings disappear and their children being used as leverage — with no idea that the evidence they need is already sitting inside the documents they've been filing and receiving for years.

Domestic violence survivors especially. Because the person who abused them in private is often the same person who performs flawlessly in a courtroom. Who knows how to sound credible. Who files more, claims more, and counts on the fact that volume alone will overwhelm whoever is sitting across from them.

The truth about what he did — to her, to their children, to the court record — was always there. It just needed something that could read all of it at once.

That's CaseBrake. It's the system I built for my wife, adapted and made available to everyone who needs it. We read your filings and find the contradictions. We map the perjury. We reconstruct the timeline. We document the pattern. We hand you everything with exact page citations, ready for court.

Or we build the full system on your computer and train you to run it yourself, forever.

Since we launched, we've worked with pro se litigants representing themselves with no attorney, and with family law attorneys who wanted a faster, more thorough way to find what's hiding in a case file. Both groups found the same thing: the contradictions were always there. They just hadn't been found yet. We're continuing to work with new clients every week — on both sides of the courtroom.

Because the lies are in the filings. They always are. They just need someone — or something — to finally read them all.