WHAT COULD YOU DO WITH
120 MINUTES?

A 94-page Supreme Court ruling lands on a Friday afternoon. The newsroom goes quiet — not because people are reading it, but because nobody knows where to start. No law degree. No time. Just caffeine and a deadline.

AmicusAI changes that. Upload any legal document — a Supreme Court opinion, an indictment, a civil complaint, a regulatory filing — and get a structured, journalist-friendly extraction in under 90 seconds.

Not a summary. A structured extraction. The ruling in plain English. Who wrote the majority. Who dissented. Key quotes with page citations. Who's affected. What to watch next. Organized the way a great producer preps a correspondent: the lede, the players, the quotes for air, and what's still missing.

Eight document types. Four analysis depths — speed mode (~30s) to deep mode (~10 min). PDFs up to 300 pages. Copy any section with one click, or download the full brief.

And when you need to go deeper: Ask AmicusAI. Once a document is parsed, you can ask it anything — What does this ruling mean for tariffs? What is Santos actually accused of? Summarize each article in plain English. It answers in plain language, grounded in the source text, with cited extracts. No hallucinations. No hunting through pages. Just ask.

AmicusAI doesn't replace editorial judgment. It does the 45 minutes of dense reading so you can spend those 45 minutes on what actually matters: deciding what the story is, who to call, and how to frame it for your audience.

Here's an open secret of national news: the best stories start local. A house fire in Phoenix. A school board meeting in Wisconsin. A wrongful conviction in Georgia. A court ruling in São Paulo. A leaked memo in Berlin. They don't announce themselves — they hide in plain sight on hundreds of station and paper websites no producer has time to monitor.

StoryScanner crawls 170+ news sources simultaneously — local TV stations, national papers (NYT, WaPo, WSJ), and 50+ international outlets across six continents — using AI to score each story's national relevance. New in this version: agentic workflows that let the tool research a thread on its own — pulling related coverage, surfacing context, and flagging what's missing, without a producer having to ask twice.

The result? Stories that would have taken hours to find (or been missed entirely) are now discovered in seconds. More importantly: we've put over 50 StoryScanner finds directly on-air — and counting.

TV scripts aren't like other writing. They're meant to be spoken — different rules, different rhythms, different pitfalls. A sentence that reads fine on paper can trip up an anchor on air. And in a newsroom, there's never time for a second pass.

CopyThat is a copy editor built for broadcast. Paste a script and it reviews every spoken line — ignoring teleprompter cues and production metadata — returning a color-coded edit with line-by-line explanations:

• Red: Spelling corrections
• Blue: Grammar and syntax fixes
• Green: Style and clarity edits (AP + broadcast tone)
• Yellow: Editorial questions — accuracy, sourcing, context

Supports AP, BBC, NPR, and Reuters style guides, plus format toggles for numbers-as-words and acronym handling. A clickable "Latest Edits" panel jumps you to flagged lines. The copy desk that never sleeps, never tires, and never misses a "their" that should be "they're."

Every producer knows this moment: a story's about to hit air, and you need a lower third. Six, maybe seven words that tell the viewer what they're watching — punchy, accurate, and ready in 30 seconds. It sounds simple until you're writing your 40th one of the day and the EP is already counting down in your ear.

Super Hero takes a block of text — a wire story, a script, a press release — and generates five broadcast-ready chyron options instantly. Different angles, different takes, all tight enough to fit a single line. Pick one, tweak it inline, copy it to your rundown. Done.

It doesn't write your show. It writes the six words you didn't have time to agonize over — so you can get back to the parts that actually need you.

(And yes — there's an easter egg. Try copying a headline that's too long for one line. Our anchorman will let you know how he feels about two-line banners. It's a personal pet peeve.)

P.S. — channel surfers: the whole site has an AITV Guide.

I wrote a LinkedIn post about how fact-checking in a newsroom is basically whac-a-mole — you catch one error and three more pop up. A MILLION that should read a BILLION. A "their" that should be "they're." Deadline looming, unreviewed copy everywhere, human eyes getting tired.

The argument: AI can kill the moles — the binary stuff, right or wrong, true or false — so journalists can spend their energy on the questions only humans can answer. Is this framing fair? Is the context changing what this quote actually means? Those aren't moles. Those are judgment calls.

Seemed like a fun concept. So I built the game.

Nine holes. Thirty seconds. Whack the editorial errors, protect the clean copy. Chiptune soundtrack included.

Here's what I've learned: you can build the most brilliant tool in the world, and it won't matter if no one uses it.

Newsrooms are full of skeptics — and they should be. Journalists question everything. Drop an AI tool in front of them and the first thought is: "What's the catch? What mistakes will it make? Will it take my job?"

My job isn't just building tools — it's building trust. Sitting with producers, showing them how the AI thinks. Being honest about what it gets wrong. Making the case that this technology handles the tasks nobody wants to do, so humans can focus on the work only humans can do.

Adoption without education is just software nobody uses. I make sure that doesn't happen.