USPTO API: Patent Data, Access Options, and AI Workflows

The USPTO Patent File Wrapper API provides access to publicly available US patent application records and associated file information through the Open Data Portal (ODP). Use it to retrieve application details, inspect prosecution records, or build a source-linked document inventory in your own application. This guide focuses on Patent File Wrapper for application-level research. ODP also provides bulk-data services for downloading datasets and processing releases locally.

What data can you retrieve?

The official Patent File Wrapper dataset listing identifies separate services for application data, documents, transactions, continuity, assignments, foreign priority, and other application information.

Application data: bibliographic information identifying an application and its associated patent record.
Documents and transactions: material for examining the application's file and processing history.
Continuity and priority: information about related applications and priority claims.
Assignments: ownership-related records for further review.

Choose the operation for the question you need to answer. To summarize an office action, retrieve and inspect the document itself; a document-list response only establishes which documents were listed. Confirm that the selected document is available for download before promising a summary.

Where do you get access?

Open the USPTO Open Data Portal. Sign in with a USPTO.gov account, complete multi-factor authentication, and supply the account information requested by the portal.
Follow Getting Started. Use the portal's API-key instructions to obtain access credentials.
Open the relevant API documentation. Start with application search, application data, or documents.
Configure a request in your API client or backend. Use the operation's documented URL, method, authentication, and input format. Keep the key in your integration settings.
Test a known application. Inspect its identifier and returned fields before adding additional retrieval or AI interpretation.

Use the query syntax examples to construct filters supported by the chosen operation. Before expanding the request, confirm that a known application's identifier and expected documents are returned.

Connect the records to an AI workflow

A practical first feature is an application document inventory:

Stage	What happens
Input	The user supplies the application identifier and a question
Retrieval	Your backend requests the relevant application records and document information
Document review	Selected document content is retrieved and supplied to the model when the question requires it
Output	The application presents a dated inventory or a summary linked to the inspected documents

For example: “List the retrieved documents for this application by date, then summarize the selected office action and cite the passages used.” This is a proposed workflow, not a completed API run. The API supplies source records; your application handles parsing, selection, and the model call. The first feature is ready for review when:

Each inventory entry retains its application and document identifiers.
The generated summary cites content actually retrieved.
Missing documents and failed requests are visible.
A reviewer can reopen the source and see when it was retrieved.

When should you use bulk data?

Bulk data fits applications that maintain a local dataset or process periodic releases. The bulk-data search API finds data products. Product information then identifies the releases and files to download. A bulk-product search returns catalog information, not a shortlist of patent documents. Plan file storage, parsing, and refreshes around the chosen product.

Expanding the research scope

Once the US application feature works, broader technical research may require international patent records, further search methods, or literature context. The Patsnap API catalog provides patent search, patent data, and literature categories for this expansion. Add the operations that answer the next research question and keep their source labels alongside the official USPTO records. The Patsnap first-request guide introduces authentication and request handling for that separate integration.

NEXT STEP · PATSNAP OPEN

Take a US record into a wider search

Start from a publication you already have and look for related disclosures beyond the USPTO set.

Explore the API reference ↗