Patsnap Patent Research

Novelty search, invention FTO review, and design-patent FTO review with staged analysis results and reports.

CommunityOnlineAPI key requiredGlobalFreeRead-only

What it can do

  • Novelty Search: Unified novelty search tool for technical solutions, invention disclosures, or claim text. If the user provides text, put it in input.text and call the tool directly; images are option
  • Fto Review: Unified FTO review tool for product or technical implementation risk review. Provide input.text and optional filters such as country, apd, legal_status, and assignee. If the user only prov
  • Design Fto: Unified design FTO tool. Provide input.images as an array of image strings. Supported formats are image URL, Data URI, or base64 content; a reachable image URL is recommended. For local cl

What data it sees

Do you need an account

An API key from the service settings is required

Novelty search, invention FTO review, and design-patent FTO review with staged analysis results and reports.

How to connect

How to connect Patsnap Patent Research to Claude

The server address to paste into Claude settings:

https://patsnap-ip-searching--openpatsnap.run.tools

  1. Open Claude (claude.ai or the desktop app).
  2. Go to Settings → Connectors.
  3. Click Add custom connector.
  4. Paste the server address copied below into Remote MCP server URL and click Add.
  5. In a chat, click + → Connectors and switch the new connector on.

Custom connectors are available on Free, Pro, Max, Team and Enterprise plans (Free is limited to one). On Team and Enterprise an organization Owner adds the connector first under Organization settings → Connectors.

Authorization

After you click Add, a sign-in window for the service opens. Sign in with your own account and approve access. Claude never sees your password.

How to connect Patsnap Patent Research to ChatGPT

The server address to paste into ChatGPT settings:

https://patsnap-ip-searching--openpatsnap.run.tools

  1. Open ChatGPT in a browser (chatgpt.com). A Plus, Pro, Business, Enterprise or Edu plan is required.
  2. Turn on developer mode once: Settings → Apps → Advanced settings → Developer mode.
  3. Open Settings → Connectors and click Create.
  4. Fill in the form: Name (anything), Description (one line about what the service does), MCP server URL (copy it below).
  5. Under Authentication choose OAuth if the service requires sign-in, otherwise None. Click Create.
  6. In a new chat open + → Apps/Connectors and enable the connector.

OpenAI has renamed this section before (Connectors → Apps/Plugins). If the label differs, search settings for "developer mode". On Business/Enterprise workspaces an admin must allow custom connectors first.

Authorization

On first use ChatGPT opens the service's sign-in window. Sign in and approve access.

How to connect Patsnap Patent Research to Cursor

The server address to paste into Cursor settings:

https://patsnap-ip-searching--openpatsnap.run.tools

Fastest: click Open in Cursor below and confirm the prompt.

Manually:

  1. In Cursor open Settings → Cursor Settings → MCP and click Add new global MCP server.
  2. Paste the JSON copied below into ~/.cursor/mcp.json (per project: .cursor/mcp.json in the repo root).
  3. Save the file. The server appears in the MCP list; authorize it there if asked.

Authorization

If the service needs sign-in, an authorize button appears next to the server in the MCP list.

Server tool list (5)

Raw names from tools/list. Only developers need these.

novelty_searchUnified novelty search tool for technical solutions, invention disclosures, or claim text. If the user provides text, put it in input.text and call the tool directly; images are optional supplemental materials. Do not ask for images when sufficient text is already provided. For local client-side image files that should be included as supplemental materials, call upload_image first to create reachable image URLs, then pass data.url values in input.images. If the user provides no technical content, ask for technical text and mention that images are optional. Agent mode rule: use execution.mode=auto unless the caller needs background execution. auto waits briefly for quick tasks before falling back to TASK_RUNNING. Use background for batch/long-running jobs or caller-managed queues. Missing or unrecognized mode defaults to auto. If the result is TASK_RUNNING, do not resubmit; call get_task using next_action until status is success, failed, cancelled, interrupted, or not_found. If a tool call is interrupted by a client, gateway, transport, or request-timeout error, do not treat it as a business failure and do not automatically resubmit; recover with get_task when a task_id was already emitted in progress or returned earlier. Responses include a presentation block for display guidance: presentation.agent tells you how to present the result; terminal detail responses may include presentation.content as displayable content to show as-is, and presentation.report_url is the full report link when available. Terminal data may still include conclusion and candidate_patents fields for clients that render structured result details. task_id is generated by MCP when omitted; caller-provided task_id values are preserved. For long-running tools/call requests or clients that want live execution logs, Clients should include tools/call.params._meta.progressToken on every tools/call request. The token belongs to the MCP request envelope, not inside arguments. Clients should consume text/
fto_reviewUnified FTO review tool for product or technical implementation risk review. Provide input.text and optional filters such as country, apd, legal_status, and assignee. If the user only provides text, do not ask for images; images are optional supplemental materials. Do not ask for images when sufficient text is already provided. For local client-side image files that should be included as supplemental materials, call upload_image first to create reachable image URLs, then pass data.url values in input.images. If the user provides no technical content, ask for a product or implementation description and mention that images are optional. Execution depth (search.mode) is separate from execution.mode. search.mode selects the FTO business pipeline: always use lite (fast risk screening, ~5 min) unless the user explicitly asks for the full pipeline; only then set search.mode=pro (multi-round iteration and report, ~30 min). execution.mode only controls tool-call orchestration. Because FTO runs are long, default execution.mode=background: submit once, then poll get_task on a fixed cadence until a terminal status. Poll every 20 seconds for lite and every 40 seconds for pro; keep polling until status is success, failed, cancelled, interrupted, or not_found, and never resubmit the task. On each poll, check data.steps for a step whose output carries a guidance field: guidance tells you how to relay to the user, in one or two natural sentences, what that step just finished and what comes next, citing the concrete values named in output.data. This keeps the user informed during the long wait. If a tool call is interrupted by a client, gateway, transport, or request-timeout error, do not treat it as a business failure and do not automatically resubmit; recover with get_task when a task_id was already emitted in progress or returned earlier. At the terminal result, the final step's guidance drives how you present the conclusion: report the overall risk summary, list candidate_patents
design_ftoUnified design FTO tool. Provide input.images as an array of image strings. Supported formats are image URL, Data URI, or base64 content; a reachable image URL is recommended. For local client-side image files, call upload_image first to create a reachable image URL, upload the binary with HTTP PUT, then pass data.url into input.images[0]. If multiple images are provided, only input.images[0] is forwarded. If the user already provides a usable product/design image, do not ask for text; input.title can be optional context. If the user provides no design content, ask for a product or design image and mention that text/title is optional. Agent mode rule: use execution.mode=auto unless the caller needs background execution. auto waits briefly for quick tasks before falling back to TASK_RUNNING. Use background for batch/long-running jobs or caller-managed queues. Missing or unrecognized mode defaults to auto. If the result is TASK_RUNNING, do not resubmit; call get_task using next_action until status is success, failed, cancelled, interrupted, or not_found. If a tool call is interrupted by a client, gateway, transport, or request-timeout error, do not treat it as a business failure and do not automatically resubmit; recover with get_task when a task_id was already emitted in progress or returned earlier. Responses include a presentation block for display guidance: presentation.agent tells you how to present the result; terminal detail responses may include presentation.content as displayable content to show as-is, and presentation.report_url is the full report link when available. Terminal data may still include final_conclusion, design_candidates, image_search, rerank, and report_export.download_url fields for clients that render structured result details. Design FTO does not forward caller-provided task_id to its backend; poll with the task_id returned in the tool response. For long-running tools/call requests or clients that want live execution logs, Clients should in
upload_imageGeneric helper tool that converts a client-side image to a URL before calling a business tool. Use upload_image when the user or client has a local image file and a business tool needs an image URL, for example before calling design_fto. This tool returns a signed upload_url plus data.url/image_url. The client should upload the image binary to upload_url with the returned method and headers, then pass data.url to the specific business tool. If the user already provides a reachable image URL, skip this helper and call the business tool directly.
get_taskUse this to fetch status for tasks created by the unified business tools or caller-provided task_id values. Use view=overview for polling and view=detail when final step outputs are needed.