Back to MCP Servers

Novelty Search

Beta
8 toolsv1.0.0
Overview

About

Turn R&D Proposals into Patent Filing Insights. Novelty Search uses the Model Context Protocol (MCP) to make Patsnap’s novelty search tools and workflow available to MCP-enabled AI applications and agents. Designed for R&D, intellectual property, and innovation management teams, it helps organizations assess the innovative value and patent potential of new technologies, product improvements, and technical solutions before investing in a patent application. Users do not need to understand patent database fields, classification systems, Boolean syntax, or complex search strategies. They simply provide an R&D proposal, invention disclosure, technical description, or supporting document. The AI application or agent can then invoke standardized tools to understand the proposed solution, extract key technical features, generate search elements, retrieve relevant patents and scientific and technical literature, compare the most relevant prior art, and perform an initial novelty and inventive-step assessment. The resulting insights help teams identify potential prior-art risks, evaluate whether an invention is worth pursuing for patent protection, and make better-informed decisions about claim drafting and patent filing strategy.

Capabilities

novelty_summary

Understand R&D proposals, invention disclosures, and related technical documents; extract the technical problem, technical solution, and technical effect; and produce a structured summary for downstream key-feature extraction and search-element generation.

Arguments
data* Requiredobject

Business parameters.

task_idstring

Task ID.

novelty_lite_submit

Submit an end-to-end Novelty Search task. After receiving technical text and optional images, the backend performs solution understanding, key-feature extraction, search-element generation, patent and scientific-literature retrieval, prior-art comparison, and preliminary assessment, then returns task_id for retrieving task progress and results.

/v2/novelty/lite_run/submit

novelty_lite_search

Run multi-channel aggregated Novelty Search retrieval across semantic, Boolean, paper, and website sources. Returns candidate patents and non-patent literature for prior-art comparison.

/v2/novelty/lite_search

novelty_lite_report_generate

Generate a Novelty Search assessment report from candidate reference documents and technical feature-comparison results, explaining distinctions from the prior art and providing preliminary novelty and inventive-step assessment, key rationale, and risk level.

/v2/novelty/lite_report_generate

novelty_lite_get

Query or poll Novelty Search task progress and results by task_id. Returns all steps or the status and results of selected steps.

/v2/novelty/lite_run/get

novelty_keywords_extract

Extract keywords, synonyms, expanded terms, and IPC classifications from technical features to generate the search elements needed for patent and scientific-literature retrieval.

/v2/novelty/keywords_extract

novelty_feature_extract

Extract and classify key technical features from the structured summary, identifying core inventive points, novelty-check points, and feature weights for retrieval and prior-art comparison.

/v2/novelty/feature_extract

novelty_feature_comparison

Compare the proposed solution’s key features with candidate patents to identify disclosed features and distinguishing points. Up to 5 patents are supported per request; if more than 5 are submitted, the service processes the first 5.

/v2/novelty/feature_comparison

What is Novelty Search MCP Server?

Turn R&D Proposals into Patent Filing Insights. Novelty Search uses the Model Context Protocol (MCP) to make Patsnap’s novelty search tools and workflow available to MCP-enabled AI applications and agents. Designed for R&D, intellectual property, and innovation management teams, it helps organizations assess the innovative value and patent potential of new technologies, product improvements, and technical solutions before investing in a patent application. Users do not need to understand patent database fields, classification systems, Boolean syntax, or complex search strategies. They simply provide an R&D proposal, invention disclosure, technical description, or supporting document. The AI application or agent can then invoke standardized tools to understand the proposed solution, extract key technical features, generate search elements, retrieve relevant patents and scientific and technical literature, compare the most relevant prior art, and perform an initial novelty and inventive-step assessment. The resulting insights help teams identify potential prior-art risks, evaluate whether an invention is worth pursuing for patent protection, and make better-informed decisions about claim drafting and patent filing strategy.

Capabilities

  • Understand R&D proposals, invention disclosures, and related technical documents; extract the technical problem, technical solution, and technical effect; and produce a structured summary for downstream key-feature extraction and search-element generation.
  • Submit an end-to-end Novelty Search task. After receiving technical text and optional images, the backend performs solution understanding, key-feature extraction, search-element generation, patent and scientific-literature retrieval, prior-art comparison, and preliminary assessment, then returns task_id for retrieving task progress and results.
  • Run multi-channel aggregated Novelty Search retrieval across semantic, Boolean, paper, and website sources. Returns candidate patents and non-patent literature for prior-art comparison.
  • Generate a Novelty Search assessment report from candidate reference documents and technical feature-comparison results, explaining distinctions from the prior art and providing preliminary novelty and inventive-step assessment, key rationale, and risk level.
  • Query or poll Novelty Search task progress and results by task_id. Returns all steps or the status and results of selected steps.
  • Extract keywords, synonyms, expanded terms, and IPC classifications from technical features to generate the search elements needed for patent and scientific-literature retrieval.
  • Extract and classify key technical features from the structured summary, identifying core inventive points, novelty-check points, and feature weights for retrieval and prior-art comparison.
  • Compare the proposed solution’s key features with candidate patents to identify disclosed features and distinguishing points. Up to 5 patents are supported per request; if more than 5 are submitted, the service processes the first 5.

How to connect

Get a Patsnap Open Platform API key, copy the MCP connection URL from the Connect panel on this page, and add the server to Cursor, Claude Desktop, or another Model Context Protocol-compatible AI agent client.

Example MCP configuration

{
  "mcpServers": {
    "novelty_search": {
      "url": "https://connect.patsnap.com/299425/mcp?apikey={your_api_key}"
    }
  }
}

Use cases

  • Prior art and novelty checks inside AI agent conversations
  • Patentability screening from technical disclosures
  • Competitive intelligence with patent and paper retrieval
  • Automated novelty report generation for R&D teams

FAQ

What is Novelty Search Lite MCP Server?
Novelty Search Lite MCP Server is a Patsnap Open Platform Model Context Protocol server that exposes novelty search workflows as MCP tools for AI agents and LLM clients.
How does Novelty Search Lite MCP Server work with AI agents?
An MCP-compatible client connects to the server URL with your Patsnap Open Platform API key, discovers available tools, and lets the model call patent retrieval, paper retrieval, and novelty evaluation tools during conversations.
What tasks can this MCP server perform?
It supports summary extraction, technical feature extraction, keyword extraction, patent and paper retrieval, novelty report generation, and novelty evaluation for innovation intelligence workflows.
How do I connect this MCP server to an agent?
Get a Patsnap Open Platform API key, open this page, copy the MCP connection URL or JSON config, and add it to Cursor, Claude Desktop, or another Model Context Protocol-compatible client.
How is Novelty Search Lite different from the full Novelty Search Agent?
Novelty Search Lite MCP Server exposes a streamlined novelty search workflow through MCP tools, while the full Novelty Search Agent on Eureka AI provides a richer interactive agent experience with broader orchestration.