MCP App Store
by KeenEthics
Seek X icon

Seek X MCP Server for ChatGPT

By NICHOLAS ALAN VANDERVELDEN

Overview

Read web pages, discover site URLs, run bounded batches and crawls, and extract structured information with source records. Seek X returns saved results with explicit progress, errors, and continuations for large pages. Article metadata includes field provenance and structured-data coverage. Browser tools support interactive workflows you request. Connect your Seek X account to access your own results and check usage. Work may wait when the server is busy. Websites may block access, content may change, and generated extractions can be incomplete or incorrect. Check material facts against returned sources. Only request sources and actions you are authorized to access or perform.

How to connect

  1. ChatGPT: Open the Seek X listing in the ChatGPT app directory and choose Connect. ChatGPT listing Seek X exposes 27 tools here.

Tools

batch

ChatGPT
Start a persistent scrape job for known URLs. Reuse the same idempotencyKey and identical arguments to recover a lost response without creating another job; changed arguments with that key return a conflict. A key is generated when omitted. Duplicate URLs within the list are separate inputs. Poll job_status even after a replay to read saved results; use job_errors for failures. Interrupted batches retain results but cannot resume.

browser_click

ChatGPT
Click one element identified by an observed selector. This can submit a form or change the external website; use only within the user-authorized workflow. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_close

ChatGPT
Use this to release the session created for the finished or cancelled browser workflow.

browser_extract

ChatGPT
Read text or an attribute from an observed selector in the existing page without changing the website. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_goto

ChatGPT
Navigate an existing isolated browser session to one HTTP(S) URL requested by the user. Replaces its current page. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_open

ChatGPT
Use this when a page needs multiple browser actions with retained state. Returns a session ID. Close the session with browser_close after the workflow.

browser_press

ChatGPT
Press one named key, optionally in an observed element. This can submit forms or change the external website; use only within the user-authorized workflow. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_preview

ChatGPT
Read a bounded current viewport without interacting with the website. Set includeSessions for a refreshed owned-session list; empty arguments list sessions. Image bytes are UI-only metadata. Do not repeat browser actions to refresh a frame.

browser_read

ChatGPT
Read the current page and observed targets without performing actions. Follow pagination.continuation for more text; a changed page is reported. Use this to inspect results, not browser_interact, so clicks and form submissions are never replayed just to read more.

browser_screen

ChatGPT
Open an embedded read-only screen for an existing Seek X interactive browser session. Refreshing the view never navigates or replays browser actions. With no ID, show your active sessions or an empty state. This is a periodically captured viewport, not a continuous video stream or extraction-job recording.

browser_scroll

ChatGPT
Scroll an existing browser page up or down. Scrolling may load additional content. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_wait

ChatGPT
Wait a bounded number of milliseconds for content in an existing browser session without sending input. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

browser_write

ChatGPT
Fill one input identified by an observed selector with supplied text. Page scripts may save the change; use only within the user-authorized workflow. Use browser_read to inspect the resulting page or continue returned text; never repeat an action just to read more.

cancel_job

ChatGPT
Use this when the user wants to stop a crawl, batch or extraction job started through Seek X. Saved results remain. Inspect the returned status: a completed job stays completed.

crawl

ChatGPT
Use this to collect pages from a site. Starts a persistent job and returns its ID; poll job_status with kind crawl. It does not wait for completion.

extract

ChatGPT
Use this to answer a question or extract JSON from supplied URLs. Starts an asynchronous, source-grounded local Agent job. Poll job_status with kind extract; local model inference can take minutes.

get_profile

ChatGPT
Return the account represented by this connection. Its opaque ID stays the same across token refresh and reconnection; email is display information. Takes no account selector and does not use credits.

job_errors

ChatGPT
Read crawl or batch failures and URLs blocked by robots.txt without downloading saved page bodies. The arrays describe observed failures/skips so far; poll job_status to determine whether the job has finished.

job_status

ChatGPT
Poll a crawl, batch or extraction job. Auto view returns compact progress while running and results when terminal. Inspect progress.outcome and skippedCount: completed jobs may have no_results, partial results, or only failures. Follow next for more saved documents or each document pagination.continuation for more text. To continue one document use limit 1.

map

ChatGPT
Use this to find URLs on a known website before choosing pages or starting a crawl.

map_result

ChatGPT
Read the next page of an existing map result. Does not re-fetch the site or charge another map. Follow next while pagination.hasMore is true.

scrape

ChatGPT
Read source text and links from a URL. Article metadata can include JSON-LD fallbacks; inspect metadata.fieldSources and metadata.structuredData for provenance and limitations. includeImages adds declared HTTP(S) image URLs, including responsive and lazy candidates; imageCoverage reports partial lists. Use render for JavaScript pages and waitFor (milliseconds, up to 10000) for delayed content. Preserve warnings: renderingFallback means static content may be incomplete. Follow pagination.continuation exactly for more text. Use includeTags/excludeTags to focus text or extract for model extraction.

search_status

ChatGPT
Read an accepted search without repeating it or charging another search. Poll after pollAfterMs while activity.status is running. On completion use activity.result; failed searches have no successful result. Progress is retained for ten minutes after completion and may be unavailable after a service restart.

status

ChatGPT
Use this to check Seek X API availability and supported features before troubleshooting a workflow.

usage

ChatGPT
Check your account credit balance, daily allowance, reset time and operation costs. Uses no credits. A local service without accounts returns operator usage instead. Do not infer a paid plan or suggest upgrades from this response.

Capabilities

InteractiveWrites

Example Prompts

Click any prompt to copy it.

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/seek-x"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/seek-x)