Engraphic Legal Library Request a demonstrationRequest demo

On your serverUpdated nightlyPatent pending

The first large language model (LLM) AI that doesn’t hallucinate!

We built a better AI and then applied it to the legal field.

Request a demonstration Read an answer

Open any citation in the answer. Each one shows the passage it rests on.

Recorded answer, condensed U.S. patent sources
Question

What is the grace period for an inventor-originated disclosure?

Several MPEP sections, including § 2153.01(a), reference the “Grace Period Inventor-Originated Disclosure Exception” and link it to AIA 35 U.S.C. § 102(b)(1)(A). None of these texts is excerpted in full.

Even where sources discuss prior art exceptions, grace periods, or disclosures, all context remains confined to U.S. statutory interpretation under the America Invents Act.

The Library does not say. It is silent on the subject.Inference

Method

Three things every response does.

Cite.

Every response is fully referenced and each citation is viewable.

Specimen

Several MPEP sections reference the “Grace Period Inventor-Originated Disclosure Exception.”

MPEP § 2190, part 7 of 11Citation viewable

Label.

Inference is clearly shown as such.

Specimen

It is silent on the subject.Inference

Abstain.

If the library doesn’t contain the requested data it says so.

Specimen
The Library does not say. The Library contains no verbatim excerpt, citation, or contextual evidence sufficient to answer what 1 U.S.C. § 1 provides on this point.Inference

Architecture

Knowledge on one side. Reasoning on the other.

Engraphic AI’s unique and patent pending tech stack separates knowledge from reasoning, which structurally eliminates hallucination.

Knowledge kept apart from reasoning Knowledge, the Library, is kept apart from reasoning, the AI. The response is fully referenced, shows inference as such, and says so when the library does not contain the requested data. KNOWLEDGE · THE LIBRARY Statutes, cases, rulings, opinions, and your own documents REASONING The AI. Kept apart from the knowledge. THE RESPONSE
  1. KnowledgeThe Library: statutes, cases, rulings, opinions, and your own documents.
  2. ReasoningThe AI. Kept apart from the knowledge.
  3. The responseFully referenced, with each citation viewable. Inference is clearly shown as such. If the library doesn’t contain the requested data it says so.

Because our AI uses better math it is more efficient, supports continuous learning and requires far less compute. This means that you don’t need an expensive data center to run it.

Patent pending. The method separates knowledge from reasoning.

On your server

Your data resides securely on your Engraphic AI server.

You can integrate your data with the AI or keep it separate.

Where it runs
On your Engraphic AI server
How it updates
Nightly updates
What it needs
You don’t need an expensive data center to run it.
A building with a server inside, enclosed by a dashed line marking where the data stays Data stays inside this line. Your Engraphic AI server

Holdings

Statutes, cases, rulings, opinions, etc.

Our Legal Library is an on-premises legal AI that functions as a smart library to look up statutes, cases, rulings, opinions, etc., with nightly updates.

The Library holdsUpdated
StatutesNightly updates
CasesNightly updates
RulingsNightly updates
OpinionsNightly updates
Your own documentsIntegrate your data with the AI or keep it separate

Request a demonstration

Request a demonstration.

We reply by email, usually within one business day.

◆ Engraphic Legal Library

legal.engraphic.io

Patent pending.