The +Drafting version of ClaimMaster 2026 introduces a slew of improvements that significantly enhance its LLM patent drafting capabilities. ClaimMaster now supports the latest OpenAI GPT-5 and Anthropic Claude LLMs through an interactive chat interface, which maintains conversation context for iterative patent drafting sessions, lets you attach various documents for analysis, batch generate claim charts, and create patent images in several formats.
Support for the latest and greatest LLM models
ClaimMaster 2026 can now directly interact with the latest LLMs via API, including GPT-5 from OpenAI and Claude 4.5 from Anthropic. It can also integrate with the most advanced local LLMs, such as gpt-oss via Ollama. In case you are pointing LLMs to your private LLM instances, we’ve also hardened encryption for the locally stored API keys using the latest Window’s cryptography libraries and simplified auto-configuration of private LLM settings for multi-user deployment.
LLM chat interface improvements
ClaimMaster 2026 features an all-new interface for interacting with the configured LLM in an intuitive chat format. The new interface lets you easily maintain a continuous chat context with the LLM for iterative patent drafting sessions – for example, you can quickly prompt the LLM to redraft or supplement its previous answers, quickly step through, review and diff LLM edits, use shorthand to reference previous answers, and much more. You can also pull-up preconfigured prompt templates populated with document information, set document workspace for RAG, and export LLM answers to your current or new Word documents, including any marked up text, such as claim amendments, bold/italicized text, and tables. To construct more complex LLM prompts, you can use a special “Prompt Crafter” interface. Notably, all LLM conversation context is permanently removed once you close the conversation window. For more information about the new LLM chat interface, click here.
Improved document processing and chart generation
The new LLM interface lets you attach and select specific pages from multiple files for LLM analysis, including PDFs, Office files, text/CSV, and images. We can also automatically download specific patents and IFW documents by patent/application number and attach them to the outgoing LLM prompts.
We’ve also added the ability to use LLMs to generate claim charts for summarizing claim limitations or comparing them to the attached prior art or other documents (e.g., for mapping infringement, etc.) . Users can select specific claims for charting and ClaimMaster will processes each claim through configurable LLM prompt templates to generate claim charts containing LLM analysis for each claim limitation.
You can even run our Batch Download tool to download multiple patents at the same time and generate LLM-based claim charts for each of those patents (as part of post-processing options):
Use LLMs to generate PNG and SVG images
The latest ClaimMaster integrates with the latest LLMs that support image generation (e.g., OpenAI) for patent figure drafting. For example, you can ask LLMs to generate images based on the attached text description or turn existing images into line drawings. One limitation of LLM image generation is that such images are not easily editable. As a possible work-around to non-editable images, you can also ask LLMs to generate drawings in SVG XML format, which can be viewed in browsers or edited in design software like Visio. SVG images also have limitations for complex images, but may work well for flowcharts or other drawings that use primarily simple geometric shapes. For more information, click here.
Tags for LLM prompt templates
ClaimMaster 2026 lets you annotate stored LLM prompts with descriptive tags for quick prompt organization and lookup. For example, you can tag a set of LLM prompts for “claim analysis”, “spec drafting”, “claim charting”, and so on.
You can then quickly sort and lookup prompts in the LLM chat window based on their tag values, allowing you to lookup specific prompts much faster:
For more information about setting up LLM prompt templates and tags, please take a look at this sections of the online manual.