HIPAA-compliant AI for personal injury law

If you're going to use AI, use the best.

Hermosa Strategy runs the best legal-reasoning models available — and tells you exactly what every query costs. The firms cutting corners on cheaper model run-rates aren't getting the answers their cases deserve.

No credit card for the trial · Cancel anytime · Your data is never used to train models

Claude Opus 4.5 Vision document reading Cited sources Per-query cost transparency Amazon Bedrock · BAA
The premium position

The best models, or none at all.

Medical records don't forgive a model that mixes up an impairment rating or invents a date. We pay for frontier intelligence because in personal injury, being exactly right is the entire job.

“Cheaper run-rates buy cheaper reasoning. We'd rather charge you honestly for a model that reads every word on every page — and gets the diagnosis, the date, and the dollar figure right the first time.”

Claude Opus 4.5

Our primary engine for analysis, extraction, and drafting. The most accurate at pulling specifics without fabricating or mixing up sources.

Primary

Claude Sonnet 4

Automatic high-availability fallback during peak demand, so your firm never sees an outage mid-case.

Fallback
Nothing hidden

We're upfront about what every query costs.

No mystery credits. No vague "tokens." Open any answer and see the exact usage and dollar cost — straight from the app, in real time.

app.hermosastrategy.com
Hermosa Strategy token usage panel showing 491,214 tokens and an estimated $7.37 cost at the Enterprise overage rate
~57K
tokens per query — your whole case folder is searched, the most relevant excerpts retrieved, and the answer generated
$0.86
approx. cost of a single overage query on Enterprise ($1.50 / 100K tokens)
$0
extra for vision analysis of scanned pages, charts & images — processed once at upload, included

Included tokens are covered by your plan. You're only ever billed for usage beyond your monthly allowance — and you can watch it accrue, query by query. You're never cut off mid-case.

How it works

It reads everything. Then it cites where it read it.

Upload a case folder and Hermosa builds a private, searchable knowledge base of it — text, tables, and the visual pages other tools skip.

01

Vision document reading

Scanned records, X-ray charts, handwritten intake forms, diagnostic images — pages that are images, not text — are run through AI vision at upload and turned into searchable, queryable content. One-time, no extra token cost.

02

A knowledge base per case

Every document is indexed into a HIPAA-eligible knowledge base, isolated to your firm. Ask a question and we retrieve only the most relevant excerpts across the entire folder — so cost stays flat no matter how big the file gets.

03

Answers that cite sources

The model only reasons over what's actually in your documents, and every claim points back to the exact source file and page. No hallucinated case law. No invented facts. Verifiable, every time.

Cited sources

Trust, but verify — instantly.

A summary is only useful if you can stand behind it in front of a judge or an adjuster. Hermosa grounds every answer in your actual records and shows its work, so you can click straight to the page it pulled from.

Because it's reading your documents — not guessing from training data — it won't fabricate a treatment date or a code that isn't there.

What was the patient's cervical spine impairment rating, and who assigned it?
The treating physician assigned a 12% whole-person impairment for the cervical spine under AMA Guides 5th Edition[1], consistent with the C5–C6 disc herniation documented on the MRI[2].
[1] IME_Report_Dr_Alvarez.pdf · p.7 [2] Cervical_MRI_2025-03-11.pdf · p.2
Real-world example

One folder. 87 documents. One question.

We loaded a full personal-injury case — five subfolders, 87 documents, roughly 4,900 pages of records, bills, imaging, and correspondence — and asked a single question. Here's what happened.

5
Subfolders
87
Documents
~4,900
Pages indexed
9
Docs cited in one answer
Hidalgo v. Coastal Freight
Medical Records38 docscited
Diagnostic Imaging14 docscited
Billing & Liens19 docscited
Police & Incident Reports6 docscited
Correspondence10 docs
Across the entire case, list every injury attributed to the collision, who diagnosed it, and the total billed treatment cost — cite each source.
The collision produced four documented injuries: a C5–C6 disc herniation[1], right rotator-cuff tear[2], post-concussive syndrome[3], and lumbar strain[4]. Billed treatment to date totals $148,210 across nine providers[5].
9 source documents 4 of 5 subfolders ~12s to answer ~58K tokens · ~$0.86
No keyword search · semantic retrieval across the whole folder Cost stays flat whether the folder is 8 docs or 800
The math

Cheaper and more exact than a paralegal.

This doesn't replace your team — it makes them superhuman. But on raw document review, the comparison isn't close.

Manual review

2–4 hrs
per 50-page medical record
  • $4,200–5,400 / month, fully loaded
  • Skims, tires, misses page 47
  • One case at a time

Hermosa Strategy

< 30 sec
per record, citing every source
  • From $149 / month for a solo attorney
  • Reads every word on every page
  • Cross-document answers, with citations
Pricing

Premium intelligence, honest pricing.

Every plan includes unlimited conversations, unlimited documents, vision reading, and full HIPAA-compliant infrastructure. Token budgets reset monthly — go over and you're billed a small, transparent overage rate, never cut off.

Starter
SOLO — 1 SEAT
$149/mo
For solo practitioners. No additional seats — upgrade to Professional as your team grows.
  • 10M tokens / mo (~175 case-file queries)
  • Unlimited conversations & documents
  • Vision document reading included
  • Overage at $2.50 / 100K tokens
  • HIPAA-compliant infrastructure
Start with Starter
Most popular
Professional
3 SEATS · +$24 / SEAT
$349/mo
Built for small and mid-size firms handling high case volume.
  • 24M tokens / mo (~420 queries)
  • Everything in Starter
  • Lower overage at $2.00 / 100K tokens
  • Priority support
Get Professional
Enterprise
10 SEATS · +$19 / SEAT
$799/mo
For large firms with heavy document review needs.
  • 57M tokens / mo (~1,000 queries)
  • Everything in Professional
  • Lowest overage at $1.50 / 100K tokens
  • Dedicated support & custom onboarding
Talk to us

A single overage query costs about $0.86–$1.43 — less than a few minutes of paralegal time.

Built for PHI

Compliance isn't a feature. It's the foundation.

HIPAA-eligible AI

Runs on Amazon Bedrock under a signed BAA. Your data is never used to train models.

Strict tenant isolation

Every document and query is row-level scoped to your firm. Nothing crosses tenant boundaries.

Full audit trail

Every access, query, and generation is logged with timestamps — append-only, for compliance.

Encrypted end to end

Encryption at rest and TLS in transit, with short-lived access to every document.

Ready when you are

Don't go halfway on the work that wins cases.

Start your trial, upload a real medical record, and see how fast — and how exactly — Hermosa reads it. No credit card required.