AI research engine
Research that searches your files and the law.
Most legal AI tools search public databases. CaseRead searches your firm’s private documents and public case law simultaneously, and cites every source.
Lite $15/mo · Pro $149/mo · Cancel anytime
Citation integrity
Every claim has a source. Period.
Cited to retrieved sources
The AI answers from documents and sources actually returned by the search. If it can’t find a source, it tells you, and on Pro and Team, any citation it can’t ground back to one is flagged before you read it as settled.
Source transparency
Every answer shows exactly which sources were searched, which were unavailable, and where gaps exist. Full coverage reporting.
Professional recommendations
When our database doesn’t cover a jurisdiction or topic, we tell you and recommend verified sources like Westlaw or Lexis for confirmation.
Clickable source links
Every citation links back to the original document, whether it’s in your firm’s vault or a public statute. One click to verify.
Dual-RAG engine
Two databases. One search. Complete answers.
One answer citing both your internal documents and controlling public authority, with clickable source links for every claim.
Public law sources
- Federal Rules of Civil Procedure
- Federal Rules of Evidence
- State statutes (all 50 states, DC & Puerto Rico) via live search
- Published case law & court opinions
- Federal regulations & administrative rules
- Federal & state constitutions
Your firm’s vault
- Client documents & contracts
- Internal memos & research
- Prior case files
- Firm templates & precedents
- Uploaded briefs & motions
- Work product & correspondence
Broad coverage. Honest about gaps: if a jurisdiction or source type is unavailable, the AI tells you.
Why it is different
Legal AI has a trust problem.
General AI tools
- Search nothing: they generate answers from training data, which leads to fabricated citations.
- Don’t show their sources, making it impossible to verify accuracy.
- Search only public law and miss your firm’s internal precedents, or only your files and miss controlling authority.
CaseRead
- Searches your firm’s private vault and public law databases simultaneously.
- Every citation links back to the original document. One click to verify.
- Any citation it can’t ground back to a source is flagged before you read it as settled.
- Shows which sources were searched, which were unavailable, and where gaps exist.
Damien Charlotin’s AI Hallucination Cases Database lists 2,095 cases worldwide in which a court or tribunal found or implied that a party relied on AI-hallucinated content (as of Sept. 28, 2026). Source: AI Hallucination Cases Database, Damien Charlotin
The AI behind it
AI you can trust.
Research-grade AI
Our primary AI engine handles complex legal reasoning, multi-factor analysis, and source synthesis. It understands legal concepts, not just keywords.
Efficient processing AI
Bulk operations like document tagging and query expansion use a specialized fast model, keeping costs low without sacrificing quality.
Your data never trains our models
CaseRead never trains on your data. AI processing runs through Anthropic’s and OpenAI’s standard commercial APIs, whose terms don’t use API inputs or outputs for model training.
Security
Private by design
Client files stay inside your firm. CaseRead gives every firm its own database schema, encrypts data in transit and at rest, and trains no AI model on what you upload.
SOC 2 aligned
Every control on our Security page is live in CaseRead today: a database schema for each firm, two-factor sign-in, a firm activity log, instant project purge, AES-256 encryption at rest and no training on your data.
Rule 1.6 and ABA Formal Opinion 512
Built for the duty of confidentiality: each firm gets its own database schema with row-level security, and nothing you upload trains any AI model.
Encrypted and US-hosted
AES-256 at rest, TLS in transit, hosted in the United States. Two-factor sign-in, which a firm admin can require.
See our Security page for details.
Common questions about AI research
See the research engine in action.
Lite is $15 a month for 20 research questions a month, up to 4 a day, with a Hallucination Shield check on any answer.
Lite $15/mo · Pro $149/mo · Cancel anytime