# US 12,602,533 B2 · Content generation with integrated autoformatting in word processors that deploy large language models

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,602,533 B2
- Document ID: US12602533B2
- Status: Granted
- Kind: B2
- Application number: 18/351,974
- Filed: 2023-07-13
- Published: 2025-01-16
- Granted: 2026-04-14
- Earliest claimed: 2023-07-13
- Assignees: Google Inc.
- Applicants: Google LLC
- Inventors: Dongeek Shin
- Web page: https://patentgrep.com/patent/US12602533B2

## CPC

- G06F 40/103: Formatting, i.e. changing of presentation of documents (automatic justification G06F40/189; automatic line break hyphenation G06F40/191)
- G06F 16/3329: Natural language query formulation
- G06F 40/166: Editing, e.g. inserting or deleting
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Techniques and systems are disclosed that perform content generation with integrated automated formatting using word processing applications that deploy large language models (LLMs). The techniques include receiving, a natural language (NL) query for a synthetic content for a document, identifying formatting rules of the document, and generating an augmented query that includes a representation of at least a portion of the NL query and a representation of the one or more formatting rules of the document. The techniques further include providing the augmented query to an LLM and updating the document with the synthetic content generated by the LLM in response to the augmented query.

## Claims

### Claim 1

A method comprising:
receiving, by a processing device, a natural language (NL) query for a synthetic content for a document displayed in a user interface of a word processing application, the document associated with metadata identifying formatting rules associated with the displayed document;
processing the NL query to identify one or more formatting rules pertinent to the NL query;
accessing, by the processing device, the metadata to select, from the formatting rules associated with the displayed document, the one or more formatting rules pertinent to the NL query;
generating, by the processing device, an augmented query that comprises:
a representation of at least a portion of the NL query, and
a representation of the one or more selected formatting rules;
providing the augmented query to a large language model (LLM); and
updating, by the processing device, the displayed document with the synthetic content generated by the LLM in response to the augmented query.

### Claim 2 (depends on 1)

The method of claim 1, wherein the NL query for the synthetic content comprises a seed content, and wherein the NL query comprises a request to auto-complete at least a portion of the displayed document.

### Claim 3 (depends on 1)

The method of claim 1, wherein the one or more formatting rules of the displayed document comprise one or more of:
an indentation rule,
a line spacing rule,
a text alignment rule,
a graphics placement rule, or
a table placement rule.

### Claim 4 (depends on 1)

The method of claim 1, wherein the formatting rules associated with the displayed document are stored in a configuration file for the displayed document.

### Claim 5 (depends on 4)

The method of claim 4, wherein the stored configuration file for the displayed document comprises at least one of:
one or more default formatting rules for a document type associated with the displayed document,
one or more formatting rules associated with a user identification (ID) of a user session associated with the NL query.

### Claim 6 (depends on 1)

The method of claim 1, wherein the representation of the one or more formatting rules of the displayed document comprises:
a natural language description of the one or more formatting rules of the displayed document, or
a first tokenized representation of the one or more formatting rules of the displayed document, the first tokenized representation generated using a tokenizing algorithm associated with the LLM.

### Claim 7 (depends on 6)

The method of claim 6, wherein the first tokenized representation of the one or more formatting rules of the displayed document comprises one or more previously stored tokens.

### Claim 8 (depends on 6)

The method of claim 6, wherein the representation of at least a portion of the NL query comprises a second tokenized representation of the portion of the NL query.

### Claim 9 (depends on 1)

The method of claim 1, further comprising, prior to providing the augmented query to the LLM:
providing a preview of the augmented query to a user session associated with the NL query, and
receiving an updated augmented query from the user session.

### Claim 10 (depends on 1)

The method of claim 1, wherein at least one of the NL query, the augmented query, or the synthetic content generated by the LLM is encrypted.

### Claim 11 (depends on 1)

The method of claim 1, wherein the displayed document is provided via a cloud-based word processing application.

### Claim 12

A system comprising:
a memory device, and
a processing device communicatively coupled to the memory device, the processing device to:
receive a natural language (NL) query for a synthetic content for a document displayed in a user interface of a word processing application, the document associated with metadata identifying formatting rules associated with the displayed document;
process the NL query to identify one or more formatting rules pertinent to the NL query;
access the metadata to select, from the formatting rules associated with the displayed document, the one or more formatting rules pertinent to the NL query;
generate an augmented query that comprises:
a representation of at least a portion of the NL query, and
a representation of the one or more selected formatting rules;
provide the augmented query to a large language model (LLM); and
update the displayed document with the synthetic content generated by the LLM in response to the augmented query.

### Claim 13 (depends on 12)

The system of claim 12, wherein the NL query for the synthetic content comprises a seed content, and wherein the NL query comprises a request to auto-complete at least a portion of the displayed document.

### Claim 14 (depends on 12)

The system of claim 12, wherein the one or more formatting rules of the displayed document comprise one or more of:
an indentation rule,
a line spacing rule,
a text alignment rule,
a graphics placement rule, or
a table placement rule.

### Claim 15 (depends on 12)

The system of claim 12, wherein the formatting rules associated with the displayed document
are stored in a configuration file for the displayed document.

### Claim 16 (depends on 15)

The system of claim 15, wherein the stored configuration file for the displayed document comprises at least one of:
one or more default formatting rules for a document type associated with the displayed document,
one or more formatting rules associated with a user identification (ID) of a user session associated with the NL query.

### Claim 17 (depends on 12)

The system of claim 12, wherein the representation of the one or more formatting rules of the displayed document comprises:
a first tokenized representation of the one or more formatting rules of the displayed document, the first tokenized representation generated using a tokenizing algorithm associated with the LLM, and
a second tokenized representation of the portion of the NL query.

### Claim 18 (depends on 12)

The system of claim 12, wherein, prior to providing the augmented query to the LLM, the processing device is to:
provide a preview of the augmented query to a user session associated with the NL query, and
receive an updated augmented query from the user session.

### Claim 19 (depends on 12)

The system of claim 12, wherein at least one of the NL query, the augmented query, or the synthetic content generated by the LLM is encrypted.

### Claim 20

A non-transitory computer-readable storage medium storing instructions that, when executed by a processing device, cause the processing device to:
receive a natural language (NL) query for a synthetic content for a document displayed in a user interface of a word processing application, the document associated with metadata identifying formatting rules associated with the displayed document;
process the NL query to identify one or more formatting rules pertinent to the NL query;
access the metadata to select, from the formatting rules associated with the displayed document, the one or more formatting rules pertinent to the NL query;
generate an augmented query that comprises:
a representation of at least a portion of the NL query, and
a representation of the one or more selected formatting rules;
provide the augmented query to a large language model (LLM); and
update the displayed document with the synthetic content generated by the LLM in response to the augmented query.