Settings

Theme

Show HN: ClearDemand – Cross-case search and drafting for injury firms

cleardemand.io

2 points by Dave_stridefuel a month ago · 2 comments · 1 min read

Reader

Hello HN,

We built ClearDemand to solve the "hallucination" problem in legal drafting. General LLMs are great at writing, but terrible at accuracy—which is a dealbreaker when citing medical evidence in a demand letter.

The Problem: Personal Injury firms spend hundreds of hours manually summarizing unstructured medical records (PDFs) to draft demand letters. If they miss a specific injury or treatment date, the settlement value drops.

The Solution: ClearDemand is a drafting platform that ingests raw case files (police reports, medical records, invoices), runs OCR, and uses RAG to generate drafts that are source-verified.

Grounded Generation: Every claim in the generated letter includes a citation linking back to the specific page in the source document.

Cross-Case Search: We index the firm’s entire history, allowing lawyers to search across past matters to find similar fact patterns or successful arguments.

Style Matching: It learns the firm’s specific tone/structure so the output doesn't sound like generic AI.

We’re offering a 14-day trial to see if the OCR pipeline holds up against your messiest scanned PDFs.

Feedback on our citation UI would be much appreciated!

Thanks.

Sharanxxxx a month ago

I wouldn't trust an LLM to write a legal document. It will invent cases.

  • Sharanxxxx a month ago

    Agreed. That’s why we built the Source-Verification layer. The model isn't allowed to generate a fact unless it can point to a specific page in the uploaded medical records. It’s designed as a drafter/summarizer for the attorney to review, not a robot lawyer that files autonomously.

Keyboard Shortcuts

j
Next item
k
Previous item
o / Enter
Open selected item
?
Show this help
Esc
Close modal / clear selection