i.
Ingest
PDF, DOCX, text, or image. Scanned pages are OCR'd, formatting is normalized, paragraphs are segmented. Any state's template.
§ How It Reads
Forty pages of legalese.
Ninety seconds of clarity.
FinePrint reads real-estate, rent, and car contracts the way a senior attorney in that practice area would — clause by clause, in context, against your side of the deal. Here is exactly how.
§ 01 The Pipeline
Six stages. One briefing.
Every contract is processed through the same pipeline regardless of state, template, or party. Consistent. Auditable. Open source.
ii.
Extract
Forty-plus standard clause types identified — financing contingency, inspection, earnest money, repair credits, title, closing, default — plus the non-standard ones.
iii.
Score
Each clause is scored one through ten against our risk taxonomy and the perspective you chose. Industry benchmarks built in.
iv.
Translate
Every flagged clause is rewritten in plain English. No "hereinafter", no "party of the first part" — just what it means and what could go wrong.
v.
Prioritize
Critical, High, Medium, Low. Each flag includes a recommended action — negotiate, accept, walk away. Sorted by what matters most.
vi.
Deliver
A one-paragraph verdict — sign, negotiate, walk. Plus a shareable PDF, copy-paste negotiation requests, and a side-by-side perspective view.
§ 02 What We Look For
Forty-plus clauses.
Every read. Every time.
Earnest money & refund windows
Inspection contingency period
Financing contingency & appraisal
Repair credit caps & conditions
Closing date & per-diem penalties
Title insurance responsibility
HOA document delivery timing
Assignment & transfer restrictions
Default & specific performance
As-is vs. warranted condition
Personal property inclusions
Lead, radon & environmental disclosures
Dispute resolution & venue
Survey & boundary contingencies
Tax & insurance proration
… and twenty-five more
§ 03 FAQ
Common questions.
Is FinePrint a lawyer?
No — and that's not what it's trying to be. FinePrint is a sharper read of your contract: it surfaces the clauses an attorney would normally flag, in plain English, so the questions you take to your attorney are the right ones. For binding decisions, please retain a licensed real-estate attorney in your state.
What states does it support?
All fifty. The clause taxonomy is built on the national universe of purchase agreements — NAR, state-association templates, attorney-drafted forms. State-specific quirks (Texas Promulgated Forms, the California RPA-CA, etc.) are recognized by their structure rather than memorized.
What data do you store?
Nothing. The contract is processed in memory and discarded as soon as the response is returned. We don't keep it, we don't log it, we don't train on it. For full sovereignty, self-host the engine — it's MIT licensed.
What AI model powers it?
Anthropic's Claude (Sonnet 4.5) for clause analysis, with a deterministic risk-scoring layer on top to normalize outputs. The prompts and rules are all in the GitHub repo. Swap models if you self-host.
Can my agent use this?
Yes — and many do. See the For Agents page. The free tier is unlimited; we're building hosted team features in the open for brokerages.
How accurate is it?
In our benchmark against attorney reviews of one hundred sample contracts, FinePrint flagged ninety-two percent of issues an attorney called out, plus another fifteen percent the attorney missed in their first pass. It is not perfect. It is also not fifteen hundred dollars. Use it to know which three things to bring to your attorney.