Claim Drafting · Patent Strategy

Patent Claim Types

Every type of patent claim — independent, dependent, product, method, Jepson, Beauregard, means-plus-function — what each one covers, when to use it, and the strategic tradeoffs.

The fundamental principle

The claims are the patent. Everything else — the specification, the drawings, the abstract — exists to support the claims. A patent with narrow claims is a narrow patent, regardless of how many pages of description it has. Claim strategy is patent strategy.

All claim types

Every patent claim type, explained

By dependency

Independent Claim

Broadest

Example

1. A computing device comprising: a processor; a memory storing instructions; and a network interface configured to transmit data...

How it works

Must stand alone without reference to any other claim. Every element must be individually recited. This is your most valuable claim — invalidation kills all dependents.

Strategy

Draft to cover the broadest real-world use case, not just your preferred embodiment. Narrow only as required to distinguish prior art.

By dependency

Dependent Claim

Narrower than parent

Example

2. The computing device of claim 1, wherein the network interface comprises a 5G modem.

How it works

Adds at least one limitation to the parent. Can depend from an independent or another dependent claim. If the parent is invalidated, the dependent is also invalid.

Strategy

Use to add specific embodiments as fallback positions. If independent claim is killed, a dependent with a key feature may survive. Also used to pre-claim specific features a competitor might argue is inventive.

By statutory category

Product / Apparatus Claim

Covers the thing itself

Example

1. An apparatus comprising: a first lever arm; a second lever arm pivotally connected to the first lever arm at a pivot point...

How it works

Infringed by making, using, selling, offering to sell, or importing into the US. Covers the product regardless of how it is used.

Strategy

Best for hardware inventions where the physical structure is the innovation. Reaches manufacturers and distributors, not just end users.

By statutory category

Method / Process Claim

Covers the steps of doing

Example

1. A method comprising: receiving, by a processor, an input signal; applying a fast Fourier transform to the input signal; and transmitting the transformed signal...

How it works

Infringed by performing the steps. Under § 271(g), also covers products made by a patented process imported into the US. Method claims raise divided infringement issues if multiple actors perform different steps.

Strategy

Combine with product claims for maximum coverage. Method claims often survive Alice more easily by tying steps to a technical process rather than an apparatus.

By statutory category

Composition of Matter Claim

Covers the composition

Example

1. A pharmaceutical composition comprising: compound X in a concentration of 5–20 mg/mL; and a pharmaceutically acceptable carrier...

How it works

Covers chemical compounds, formulations, biological compositions, materials. Particularly important in pharma and biotech.

Strategy

The gold standard for pharma — composition claims cover the molecule regardless of what it is used for, blocking generic competition for the full 20-year term.

By subject matter

Beauregard Claim

Software stored on medium

Example

1. A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to: receive user input; generate a recommendation based on the input; and display the recommendation...

How it works

'Non-transitory' excludes transitory signals (EM waves). These claims reach software distributors separately from method claims targeting users.

Strategy

Always include alongside method claims for software patents. Lets you pursue software distributors (app stores, enterprise software vendors) who might argue they don't 'perform' method steps.

By format

Jepson Claim

Improvement only

Example

1. In a vehicle pedal assembly comprising a base and an adjustable pedal arm, the improvement comprising: an electronic sensor attached to the pedal arm...

How it works

Preamble is admitted prior art — the characterizing clause defines what is claimed as new. Rarely used because the prior art admission can hurt in prosecution and litigation.

Strategy

Use only when the prior art context is absolutely clear and well-established, and when you want to explicitly carve the improvement from the prior art. Most prosecutors avoid this format.

By format (§ 112(f))

Means-Plus-Function Claim

Corresponding structure + equivalents

Example

1. A device comprising: means for detecting an input signal; and means for converting the input signal into an output format...

How it works

Automatically limited to structures in the specification plus structural equivalents (not the broad doctrine of equivalents). Specification must describe the corresponding structure explicitly.

Strategy

Generally avoid unless you have a strong reason. The apparent breadth is deceptive — the specification's disclosure controls scope. See our full means-plus-function guide.

Fees and limits

Claim counts and USPTO fees

USPTO fees apply beyond the base count of 20 total claims and 3 independent claims. Fees scale by entity size (large / small / micro).

Base total claims

No excess claim fee

Base independent claims

No excess independent claim fee

Excess claim fee (large)

$100/claim

Each claim above 20

Excess independent fee (large)

$460/claim

Each independent above 3

Multiple dependent claims

Surcharge

A claim depending from >1 parent claim triggers an additional fee

Maximum claims

No limit

Fees apply for each claim above 20 total / 3 independent

FAQ

Claim type questions

What is the difference between independent and dependent patent claims?

An independent claim stands alone — it does not reference any other claim and defines the broadest scope of protection by itself. A dependent claim refers back to a parent claim ('claim 1' or 'claims 1–3') and adds one or more additional limitations, narrowing the scope. Every dependent claim includes all the limitations of its parent claim plus the added limitation. If an independent claim is invalidated, all claims that depend from it are also invalid. If a dependent claim is found not infringed (because the accused product lacks the additional limitation), the broader independent claim may still be infringed. A well-drafted application typically has 1–3 broad independent claims with a cascade of dependent claims as fallback positions.

What is a Jepson claim and when should you use one?

A Jepson claim is an improvement claim with a distinctive two-part structure: a preamble that describes the prior art context ('In a device comprising [prior art elements]...') and a characterizing clause that describes the improvement ('the improvement comprising...'). By using a Jepson format, the patentee effectively admits that the elements in the preamble are in the prior art — this admission can limit claim scope and be used against the patentee in prosecution and litigation. Jepson claims are rarely used today because most practitioners prefer to avoid any admission of prior art scope. The Jepson format may be useful when the prior art is well-established and the improvement is clearly distinct, and when the applicant wants to signal precisely what is new.

What is a Beauregard claim?

A Beauregard claim is a type of software patent claim directed to a non-transitory computer-readable medium (storage medium) that stores instructions which, when executed by a processor, cause the processor to perform certain steps. The name comes from In re Beauregard (Fed. Cir. 1995), which held that software on a disk is a manufacture and thus patent-eligible. A Beauregard claim looks like: 'A non-transitory computer-readable medium storing instructions that, when executed by a processor, cause the processor to: [functional steps].' The 'non-transitory' qualifier was added after Nuijten (Fed. Cir. 2007) to exclude purely transitory signals (electromagnetic waves) which are not patent-eligible subject matter. Beauregard claims let software patent owners reach distributors of software separately from direct implementers.

What is the difference between product claims and method claims?

A product claim (also called an apparatus claim) covers a thing — a physical device, system, composition, or manufactured article. A method claim (process claim) covers a sequence of steps for doing something. The same invention can be claimed as both a product and a method, and a well-drafted application typically includes both types. Product claims are infringed by making, using, selling, or importing the product. Method claims are infringed by performing the steps. An important practical difference: under 35 U.S.C. § 271(g), a product made by a patented process is also infringing when imported into or sold in the US — this extends method claim protection to products produced overseas by the patented process. Method claims can also reach actors who perform steps across different entities (divided infringement), which can make proving infringement more complex.

How many claims can a patent application have?

A patent application can have any number of claims, but USPTO fees apply beyond the first 20 total claims and the first 3 independent claims. As of 2024, the excess claim fee is $100 per claim beyond 20 (large entity) and the excess independent claim fee is $460 per independent claim beyond 3 (large entity). Small entity and micro-entity applicants pay 60% and 40% of large entity fees respectively. There is no statutory maximum number of claims. Complex inventions (semiconductors, pharmaceuticals, mechanical systems) often have 30–100 claims covering multiple embodiments, use cases, and fallback positions. The USPTO tracks multiple dependent claims (a dependent claim that depends from more than one parent claim) separately and charges higher fees for each.