Not just another AI patent drafting platform.

InventGenie is not a law firm and does not provide legal advice.

01

It helps you invent — not just write it up.

your one
idea
40 inventive
principles
alternative
embodiments
you pick
what stays

Most AI tools reformat what you already thought of. The Innovation Engine works the contradiction in your design, proposing alternatives grouped by inventive principle — the method engineers use to break a design deadlock. You choose what stays, so the disclosure widens before a claim is written.

02

Prior art is checked while you draft, not after you file.

10M+
patents, searched semantically
while you draft

A search run after the draft is finished only tells you what you already lost. InventGenie searches while the claims are still forming, retrieving conceptual overlap that keyword search cannot find. Every hit is scored, and each suggested amendment is scored for novelty gain — so you meet the blocking reference while amending is still free.

03

It understands your invention, not just your disclosure.

You edit it directly
CREATE_NODEComponent · "piezo actuator"
UPDATE_EDGEweight 0.4 → 0.9
DELETE_NODEmisparsed Constraint

Before drafting begins, your disclosure is read as a structure, not a document — decomposed into a typed knowledge graph of components, functions, constraints and the relations between them. Everything downstream is generated against that graph, which is why the sections do not contradict each other. And the graph is yours to edit — change how your invention is understood, and every section downstream follows.

Built by an inventor,
for inventors.

Dr. Simon Park is a professor of mechanical and manufacturing engineering at the University of Calgary, founder of five deep-tech companies, and a named inventor on granted US patents in nanocomposite sensing. InventGenie is the drafting process he wanted for his own filings.

Dr. Simon Park
Dr. Simon Park
Founder, InventGenie
LinkedIn

The founders who filed with it

They brought us a rough disclosure. They left with a draft.

★★★★★
HitchMail
“We came in with a rough idea and a whiteboard photo. What came out actually read like a patent — the prior art pass surfaced two overlaps we hadn't thought about, and we reshaped the claims before spending a dollar on outside counsel. It compressed weeks into an afternoon.”
Stefan Trischuk
Stefan Trischuk
HitchMail
★★★★★
Rukbe Tech Inc
“InventGenie made the whole process simple and fast. I was impressed by how detailed the draft was compared to anything I've tried before. It gave me confidence that my idea was being protected the right way.”
JP Rukebesha
JP Rukebesha
Co-Founder & CEO, Rukbe Tech Inc
★★★★★
Expense Trend
“The system gave me clear, high-quality patent drafts in a fraction of the time I expected. I've tried using general AI tools before, but nothing comes close to the accuracy and structure InventGenie provides.”
Faraz Bhojani
Faraz Bhojani
CEO, Expense Trend

Disclosure in, filing-ready draft out

Five stages, each feeding the next from the same invention model.

1

Disclosure

Upload a document or fill the invention form. Rough is fine.

2

Innovation Engine

Alternative embodiments by inventive principle. You pick what to keep.

3

Knowledge Graph

The typed ontology is built — and you can correct it.

4

Claims + Prior Art

Claim tree generated, then tested against the prior art corpus.

5

Validate & Export

Scored section by section, then exported as DOCX.

Innovation Engine Solutions
Showing solutions for: Hybrid Immersion-Pressurized Stovetop Coffee Brewer
Generate New Solutions
Retrigger Innovation Engine
Trigger Knowledge Graph
Generate Knowledge Graph, Claims and Patentability
No solutions selected
Cheap short-living objects
2 solutions available
Overview:

Create highly ephemeral compute units that are instantiated on-demand for short-lived tasks, and immediately discarded after completion; orchestration and redundancy provide reliability while the disposability optimizes latency and throughput.

Read Details
Select Solution
Overview:

Use ephemeral data artifacts and single-use inference/pipeline stages that are regenerated frequently to deliver low-latency results while preserving correctness through versioned snapshots and robust fallback mechanisms.

Read Details
Select Solution
Hide Labels Edit Graph Retrigger Knowledge Graph
Proceed to Claims
📊 Knowledge Graph
Graph Statistics
Total Nodes:13
Total Edges:58
Avg Degree:8.92
Node Types
artifact:7
concept:4
document:2
Node Categories
People & Organizations
Concepts & Methods
Documents & Data
Places & Institutions
Communities & Groups
Other
Included (3)
Figure 1
Figure 2
Figure 3
Add figure
Close Figure 3 Generate Download Descriptions 0
FIG. 1 40 42 20
Figure description
FIG. 2 is a schematic view of the bracket upon the track with the belt removed.

Everything in one workflow

No stitching together a chatbot, a search tool and a draftsman.

Claims + specification

Claim tree, abstract, background and detailed description from one connected record.

Prior art before you file

10M+ USPTO, EPO and CIPO records searched semantically, amendments scored for novelty gain.

Figures with numerals

Sketches become labelled figures with lead lines and numerals that match the spec text.

Confidential by default

Encrypted, and never used for model training — a public disclosure can cost you the filing.

An editable invention model

See how the AI understood your invention and fix it directly, instead of re-prompting.

Export and file

Validated draft out as DOCX, ready to submit through USPTO Patent Center yourself.

A provisional starts a 12-month clock.

Start it with a disclosure that will still hold up when you convert — or draft the full application here too.

Patent drafting insights, occasionally

Claim-writing tips, AI drafting practice, and product updates. No spam.