"Writing about MCP" matching MCP servers:

Enables reading and writing Chrome bookmarks (folders, URLs, search, move, remove) through MCP clients like Claude Code, using a companion Chrome extension for secure local communication.

An MCP server that uses Playwright to scrape Israeli weather forecasts from weather2day.co.il by automating a real browser, enabling an LLM to answer questions about current conditions and hourly forecasts for Israeli cities.

Ghost Hands gives agents governed, recorded, replayable hands on the web through a zero-dependency Python CDP stack, classifying every action before it runs and writing it to a provenance trail.

Enables writing Naver blog posts with rich formatting like images, tables, formulas, and places, converting Markdown to drafts and handling publishing via MCP tools.

Enables importing grades from an Excel or CSV file into Navarra's Educa portal by driving a local browser to match student names and fill in the gradebook, showing a preview and writing values on screen before saving only after explicit confirmation, while never closing, signing, or publishing an evaluation.

Enables AI agents to generate SEO-optimized CSDN blog drafts, refine them against a writing style guide, run SEO self-checks, and optionally publish them to CSDN's draft box via browser automation with rate limiting and a human review gate. Also supports session-based login management, local draft library operations, and reading article statistics through CSDN's public API.

MCP server that fetches web pages and uses local Qwen models to answer questions about their content, with site-scoped search and map-reduce chunking for long pages.

Fetches web pages, converts HTML to markdown, and uses Claude to answer questions about the content.

Lets an AI agent read public TikTok data with no API keys or developer account: search videos and hashtags, fetch profiles, recent videos and comments, discover creators posting about a topic, and transcribe videos through your own OpenAI-compatible speech-to-text endpoint. All popular MCP clients can launch it via uvx, and every tool returns structured output as a read-only operation.

Enables coding agents to examine a rendered web page by cross-checking what the browser reports about the DOM/CSS against what is actually painted in pixels, pinpointing which element is broken and the evidence why. Reports visual problems without requiring a baseline, and is designed to grow into design-token extraction and accessibility/UX audits built on Playwright.

Enables AI assistants to build and simulate Tinkercad Circuits through a real Chrome session, including placing parts, wiring pins, writing Arduino code, running simulations, and reading serial output.

Lets an AI assistant read and fill an Innoslate project through your own logged-in browser session, surveying the project, reading your documents and guides, and proposing a plan before writing entities, trace links, artifacts and reports after a dry run. It works on the free tier where API keys are refused, since every call goes through the browser instead.

Enables AI agents to work inside a user's signed-in Buildertrend account, reading job schedules, identifying late tasks, writing daily logs, and producing progress updates without requiring a Buildertrend API key.

Enables Claude and other MCP clients to search finn.no and create, manage, and publish Torget ads from photos — preparing images (EXIF/GPS stripping, resize, HEIC), writing titles and Norwegian descriptions, suggesting category and market price, checking for duplicates, and publishing only after explicit confirmation. Read-only search, listing details, saved-search diffs, and management of your own ads are also supported through a locally stored logged-in session.

An unofficial MCP server that lets MCP clients like Codex operate the ChatGPT web interface through a persistent browser, enabling prompt writing, file uploads, message sending, and reading complete responses without using the ChatGPT API.

Enables AI assistants to remotely control TradingView charts in Chrome, including reading charts, writing and fixing Pine Script, managing watchlists and alerts, and capturing screenshots, all through natural language commands.

Enables Claude to delegate entire browser tasks to TypeSafe's Jev, which performs each click/type/select decision in about 100 ms, allowing fast automated web interactions while Claude handles planning and verification.

Enables Claude Code and Gemini CLI to see your macOS screen, answering questions about the file, error, UI, database table, or web page you're looking at. Captures screenshots, reads database schemas, and controls Chrome/Brave, presenting answers as on-screen highlights or spoken bubbles above your cursor.

Enables LLMs to autonomously retrieve and explore web content by fetching pages and recursively following links to a specified depth, particularly useful for learning about topics from documentation.

An MCP server that allows users to check if a website is experiencing downtime by querying isitdownrightnow.com, providing status information and details about recent downtime events.

Enables an AI agent to drive a real LinkedIn account through a persistent browser session, covering full profile editing, people and content interactions (searching, connecting, messaging, posting, reacting), and job search with step-by-step Easy Apply submissions. It reads data through LinkedIn's internal Voyager API while writing through the actual UI, with a saved answer bank so repeated application questions are resolved automatically.

MCP server for the Chocodata web scraping API. Enables agents to scrape live product prices, search results, and social profiles without writing glue code.

Drive Clearcote, a Chromium build with its fingerprint protection patched into the engine, from any MCP client: about 20 tools to navigate, read, extract, click, fill, screenshot and keep sessions, plus a CDP endpoint that hands the same browser to Playwright or Puppeteer. The server lives in the repo's mcp/ folder (npx clearcote-mcp or pip install clearcote-mcp).

Exposes a governed, OAuth-gated suite of 38 tools that let web AI clients work directly on your machine: reading, writing, and editing files, running whitelisted shell commands, searching, Git, SQLite, browser automation, DevTools debugging, background tasks, localhost/LAN fetches, clipboard, notifications, port management, and Postman control.

Enables an LLM to drive a real browser by declaring one outcome at a time, while Jev chooses the element, action, and value and Playwright executes it, returning a clear status such as done, blocked, or needs confirmation. Also exposes tools to open pages, ask yes/no or multiple-choice questions about a page, read question-relevant page text, capture snapshots and screenshots, act on elements directly, and invoke a page's own WebMCP tools.

Enables AI coding agents to drive a real browser like a person — crawling every page and role, running real user journeys, and auditing controls a DOM-based test misses — so UI, accessibility and permission bugs are found and only the screens affected by recent changes are re-tested. Runs are script-driven and read-only by default, so testing costs no model tokens and nothing about the app is sent to an AI.

Lets AI agents use Paper's design tools — writing HTML, capturing screenshots, reading JSX, and creating files — by driving the Paper editor in a headless browser instead of requiring the desktop app. It serves these tools over stdio to any MCP client, such as Claude Code, Cursor, or Codex.

Enables AI agents to pull deterministic, traceable numbers about Adobe Stock supply, seasonality, keyword demand and the contributor's own portfolio and Adobe accept/reject outcomes. It also supports upload staging and metadata application on the portal for a verified plan, while never clicking Submit or Delete.

Enables AI agents to query live data from wemakedevs.org — hackathons, timelines, rules, resource hubs, blogs, meetups, scholarships, careers and registration status — through 16 typed tools that return schema-validated structured JSON. Every call reads the site live via TinyFish, so agents can answer questions about deadlines, prizes and open registration without canned data.

Enables AI agents to discover postings from watched company boards ranked by keyword fit, tailor and fact-check resumes against a source-cited profile, render ATS-readable PDFs, submit applications through the user's own browser, and track replies and follow-ups. All ranking, claim checking and email classification run as plain code, so LLM tokens are spent only on writing.

Enables coding agents to search and retrieve real production websites, their design systems, and reference components over MCP, so they can study design patterns before writing UI.

Enables local MCP clients to answer questions about the current browser page by capturing DOM context, routing them to fast or deep OpenAI models via a local model router, and optionally filling forms and submitting answers.

Enables AI agents to browse, click, type, screenshot, and extract data from web pages without writing CSS selectors, using numbered element refs. It also supports office automation such as bulk form filling, table extraction, downloads, PDF archiving, and page monitoring.

Enables agents to fetch public URLs as noise-free Markdown (via Jina Reader with a local HTML fallback) protected by SSRF and indirect prompt-injection defenses, while capturing each article's hook, heading rhythm, and sentence-ending style as a reusable writing recipe in a local DuckDB style vault so that recipe can later drive automated manuscript generation.

Give AI agents visual + structured feedback about live web UIs — and give humans deterministic demo reels while you're at it.

Enables Claude to search, read, summarize, and ask about your own Microsoft Teams chats stored locally in SQLite, with tools for listing chats, reading messages, fetching images/links, syncing full history through a logged-in browser session, and suggesting replies. It works with the new Teams macOS cache and does not require Microsoft Graph or admin consent.

Enables ChatGPT to work on a local machine through an MCP connection by inspecting projects, editing files, running bounded terminal and Git commands, driving Chrome or Edge, and reading or writing an already-connected Excel workbook. It invokes the Codex model explicitly only when a task genuinely needs it, so ordinary local tool actions consume no Codex model usage.

Enables agents to extract structured data from web pages by supplying a URL and a natural-language prompt describing the desired fields. This lets assistants retrieve live site content, such as lists of videos, without writing custom scrapers.

Enables AI agents and developers to safely interact with an authenticated local LinkedIn browser session through the Chrome DevTools Protocol, supporting profile inspection, headline and About updates, project and experience additions, post creation, and screenshots without storing credentials.

Enables AI assistants to control a logged-in TradingView chart in Chrome, including writing and compiling Pine Script indicators, adding them to charts, changing symbols and timeframes, capturing screenshots, and retrieving strategy tester results.