Browser automation and web scraping via Playwright. Navigate, click, fill forms, take screenshots, and extract structured data from any website.

Web scraping and crawling with LLM-ready output. Scrape, crawl, map, and extract structured data from any URL. Returns clean markdown.

Web and local search via Brave Search API. Privacy-respecting search with real-time results, news, and image search capabilities.

Semantic and neural web search via Exa. Find exactly what you mean, not just keyword matches. Ideal for research agents that need high-quality results.

HTTP fetch and web content extraction for agents. Retrieve any URL as markdown, HTML, or raw text. Essential for agents that need to read the web.

Official MCP server for the Google Maps Platform API, providing geocoding, directions, places search, and distance matrix tools.

MCP server for Tavily's AI-optimized search API, purpose-built for agent web research with structured, LLM-friendly results.

Official AWS Labs MCP server for searching and retrieving AWS documentation, best practices, and getting-started guides.

Read and write content in Contentful CMS. Query entries, update fields, publish content, and manage assets from AI agent workflows.

Interact with Sanity Studio datasets. Query documents with GROQ, create and update content, and manage assets in Sanity-powered projects from AI agents.

Run web scraping and automation Actors from the Apify platform. Access a library of 3,000+ ready-made scrapers for social media, e-commerce, search engines, and more.

Cloud browser infrastructure for AI agents. Run headless Chromium at scale with session management, stealth mode, CAPTCHA handling, and streaming. Purpose-built for agent web automation.

Puppeteer browser automation via MCP. Navigate pages, click, fill forms, take screenshots, and extract content using headless Chrome.

Stagehand AI-native browser automation via MCP. Natural language browser control — act, extract, observe — powered by Browserbase infrastructure.

Manage Webflow sites, CMS collections, and pages via MCP. Agents can create and update CMS items, publish changes, and read site structure — enabling AI-driven content operations on Webflow.

Manage WordPress sites via MCP using the REST API. Create posts, update pages, manage media, handle comments, and query site data — giving agents full CMS control over WordPress properties.

Official MCP server for Amazon Advertising with 50+ tools covering Sponsored Products, Sponsored Brands, Sponsored Display, DSP, and Amazon Marketing Cloud. Full campaign management through AI agents.

Official Microsoft Learn MCP server and CLI. Powers LLMs and AI agents with real-time, trusted Microsoft documentation and code samples. Only accesses official first-party Microsoft docs.

MCP server for Strapi — the leading open-source headless CMS. Create, read, update, and delete content types, manage media, and integrate AI agents with Strapi-powered content workflows.

Official Bright Data web data infrastructure MCP server. Routes scraping requests through 400M+ residential IPs for anti-blocking. Provides SERP API, scraping browser, and structured datasets for high-difficulty web data extraction.

Jina AI Reader MCP server. Converts any URL to clean, LLM-optimized markdown with a single tool call. Ideal for RAG pipelines, content summarization, and feeding web content into agent workflows without scraper configuration.

Official MCP server for TikTok Ads Manager. Enables AI agents to create, manage, and optimize ad campaigns, set bids, adjust budgets, modify targeting, upload creatives, and monitor performance without manual dashboard intervention.

Open-source MCP server for Google Ads platform. Enables AI agents to create and manage ad campaigns, adjust bids, configure targeting, monitor performance metrics, and generate reports across Search, Display, and YouTube ad networks.

MCP server for Optimizely CMS providing AI access to GraphQL API and Content Management API with dynamic schema introspection and visual builder support.

MCP server integrating Naver Search API and DataLab API for Korean web search, news, blogs, shopping, and trend data. Supports category-based search, trend analysis, and shopping insights with comprehensive access to Korea's dominant search ecosystem.

MCP server from Orderly Network enabling AI agents to build, launch, and manage perpetual futures decentralized exchanges across 15+ blockchain networks. Consolidates API documentation, SDK patterns, contract addresses, and operational workflows into a queryable context layer that integrates with Orderly One, the project's no-code DEX launcher supporting 110+ trading assets.

Experimental MCP server from Mozilla providing AI assistants with access to MDN Web Docs search, documentation content, and Browser Compatibility Data (BCD). Enables AI coding tools to query authoritative web platform documentation, check cross-browser API compatibility, and reference current web standards directly through the Model Context Protocol.

Open-source MCP server that enables AI assistants to interact with Squarespace sites through 67 tools covering all major platform features. Built with TypeScript and the official MCP SDK for type-safe reliable access to Squarespace commerce operations including orders, products, inventory, and transactions. Also covers customer and subscriber profiles, webhook management, page and blog management, form submissions, and analytics. Works with Claude Desktop, Claude Code, and any MCP-compatible client for full Squarespace site management through natural language.

Open-source MCP server from GravityKit that enables AI agents to edit WordPress posts at the Gutenberg block level without breaking block structure. Provides block-aware editing capabilities that understand WordPress block markup, allowing AI tools to insert, modify, reorder, and delete individual blocks while preserving the overall document structure. Solves the problem of AI agents corrupting block serialization when editing WordPress content as raw HTML. Works with Claude Desktop, Cursor, and other MCP clients. Community-maintained and open source.

MCP server for complete management of PrestaShop e-commerce stores. Provides AI agents with tools to manage products, categories, customers, orders, modules, cache, themes, and navigation menus. Supports prompt templates and resources for contextual AI interactions. Includes token-based and guarded no-auth modes. Works with Claude Desktop, Cursor, and any MCP-compatible client. Open source on GitHub.

Real-time retail product data MCP server providing inventory and pricing from 80+ major US retailers including Walmart, Target, Home Depot, and Nike. Aggregates product catalog, availability, and pricing data through a unified API and MCP endpoint. Designed for shopping agents, price comparison tools, gift finders, and reseller sourcing workflows. Commercial with tiered pricing. Created by LemmeBuyIt.

Cross-border e-commerce product catalog MCP server for AI agents. Aggregates 11M+ products from Singapore, Southeast Asia, and US markets including Shopee, Lazada, Amazon SG, Courts, and Challenger. Provides search, price comparison, product details, affiliate links, and catalog browsing tools. Semantic search with filtering by category, price, and region. Free developer beta access with rate limits. Open source on GitHub.

Global shipping and logistics MCP server that connects AI agents to 550+ couriers across 200+ countries. Provides tools for comparing shipping rates, generating labels, scheduling carrier pickups, calculating import taxes and duties, retrieving tracking information, and pulling shipping analytics. First cross-border shipping MCP with real-time duty calculation. Free for all Easyship accounts. Open source on GitHub and PyPI. Created by Easyship.

Subscription commerce MCP server that plugs enterprise subscription data into AI tools for analytics, payment recovery, and retention workflows. Part of Ordergroove's Autonomous Subscriptions platform with AI agents for subscriber signal analysis, churn prediction, and recurring revenue optimization. Works with Claude, ChatGPT, and MCP-compatible clients. Commercial with managed hosting. Created by Ordergroove.

MCP server from DebugBear providing AI assistants with access to web performance monitoring and testing capabilities. Create projects, configure page monitoring, run performance tests, analyze Core Web Vitals, and automate bulk changes. Supports the same functionality as the existing DebugBear API. Works with Claude, ChatGPT, Gemini, and Cursor.

Official MCP server from Visualping, the website change monitoring platform used by 85% of Fortune 500 companies, launched in public beta July 1, 2026. Gives AI coding agents the ability to create, edit, and view web page monitors in plain language and get clear summaries of what changed each time a page updates. Available on every Visualping plan including the free tier. Compatible with Claude, ChatGPT, Cursor, and other MCP clients.

MCP server from Tiiny Host that enables AI assistants to publish, update, and delete hosted web projects directly from a chat conversation. Supports HTML files, ZIP archives, PDFs, PHP apps, and over 100 other file types. Users connect via Claude's custom connector at mcp.tiiny.host and get a live tiiny.site URL without leaving the chat window. Available on all Tiiny Host plans including the free tier. Paid plans range from $5/mo to $74/mo with custom domains and additional storage. Launched July 13, 2026.

MCP server from Parallel AI that gives LLMs and AI agents instant access to real-time web knowledge through a plug-and-play search API. Exposes the Parallel Search API as an MCP tool that AI assistants can call to retrieve current web information, eliminating the need for custom search integrations. Designed for developers building AI applications that need grounded, up-to-date answers from the web. Supports simple configuration and works with any MCP-compatible client.

MCP server and web search API from Dogpile, the metasearch engine operated by System1 (NYSE: SST) that has been aggregating results across multiple search indexes since 1996. Designed specifically for AI agents that need broader web coverage than any single search index provides. Every query is classified and routed to the best backend automatically, returning ranked results, knowledge graph data and People Also Ask questions. Includes a Deep mode that breaks complex queries into multiple searches and deduplicates results. Installs in Claude Code, Cursor or any MCP-compatible client in under two minutes. Pricing starts at $2 per 1,000 calls with a free monthly credit and no credit card required. Announced July 23, 2026.

Hosted MCP server for AlertChecker that lets ChatGPT, Claude and other MCP clients create AI-monitored web alerts from plain-English statements. Users can ask to be notified when a condition becomes true, such as a price drop, stock availability, weather condition, stock-price threshold or earthquake event. Uses streamable HTTP with OAuth 2.0, PKCE and dynamic client registration. Public GitHub repository created August 9, 2026.

Educational MCP server for safe geo-targeted web access through DataImpulse residential proxies. Instead of setting global HTTP_PROXY or HTTPS_PROXY variables, it creates a per-request proxy agent from DataImpulse credentials so other terminals, editors, package managers and MCP servers remain unaffected. Public GitHub repository created August 12, 2026.

MCP server, CLI and TypeScript library for web research in one tool call: search, fetch full pages, chunk, rank and return cited passages. It works with zero API keys using lightweight local retrieval, with optional embeddings from local Transformers or external providers when configured. Public GitHub repository created August 18, 2026.

AI-first browser exposed through a local MCP server that returns compact goal-based page views instead of full screenshots or raw page source on every turn. It uses Playwright and lazy image retrieval to lower browser-agent context costs while still supporting semantic action references, article reading, form interactions and computed-style inspection. Public GitHub repository created August 19, 2026.

Privacy-first, self-hostable MCP web search and content extraction gateway implementing the 2026-07-28 stateless MCP core with stdio and HTTP modes. It provides multi-engine search, raw fetch, HTML-to-Markdown extraction and SSRF hardening with DNS and redirect revalidation before requests reach private networks. Public GitHub repository created August 15, 2026.

Buyer-run, ad-neutral shopping-agent MCP server that compares products against a user brief, produces deterministic ranked shortlists with reasons, records which stores could and could not be searched, and requires separate approval before checkout. It keeps affiliate and seller payment signals out of ranking and creates a local audit trail for agent-mediated buying. Public GitHub repository created August 17, 2026.

Multi-engine web search MCP server for Cursor, Codex, Claude Code, Grok Build and Antigravity. It runs free engines such as Bing, DuckDuckGo, Yahoo and Exa-free in parallel, can add keyed Tavily, Brave or Exa providers, and includes X/Twitter fallback, Jina page fetch and deep-research rounds. Public GitHub repository created August 17, 2026.

Fork of Bright Data's MCP server that adds a Scraper Studio lifecycle for public-site scraper creation, breakage detection, repair and verification. It keeps Bright Data API access behind a local stdio package and exposes the expanded scraper workflow to Claude Code, Claude Desktop, Cursor and other MCP clients. Public GitHub repository created August 19, 2026.

Hosted Google Search and SERP MCP server from HasData. It exposes read-only tools for live organic results, AI Overview, People Also Ask, AI Mode, news, shopping, product detail and short-video result data without needing a Google Cloud project. Public GitHub repository created August 21, 2026.

CLI, stdio MCP server and hosted Streamable HTTP MCP endpoint for generative-engine-optimization checks. It scores crawler access, structure and citability for ChatGPT, Perplexity, Claude and Google AI, including robots.txt checks for AI crawlers and content quotability analysis. Public GitHub repository created August 25, 2026.

Official BestPrice.gr Streamable HTTP MCP endpoint for live product search, offer comparison and price-history lookups. The read-only public service exposes search_products, compare_offers and get_price_history without requiring a BestPrice account or API key. Public GitHub repository created August 26, 2026.

Hosted Streamable HTTP MCP server from HasData for read-only Booking.com search and property-detail retrieval. Agents can search stays by destination, dates and filters, then fetch full public property data as structured JSON without a Booking.com account or local scraper. Public GitHub repository created August 24, 2026.

Hosted Streamable HTTP MCP server from HasData for Zillow listing search and property details. It exposes read-only tools for for-sale, for-rent and sold listings with rich filters, returning public listing data as structured JSON without a Zillow account. Public GitHub repository created August 24, 2026.

Hosted Streamable HTTP MCP server from HasData for Google Flights-style itinerary search. It returns one-way, round-trip and multi-city flight options with fares, legs, carbon emissions and price-history signals as structured JSON. Public GitHub repository created August 24, 2026.

Hosted Streamable HTTP MCP server from HasData for read-only Airbnb stay search and listing details. It searches public listings by location and dates, then returns full listing data as structured JSON without an Airbnb developer account or partner approval. Public GitHub repository created August 24, 2026.

Hosted Streamable HTTP MCP server from HasData for Google Trends data. It exposes a read-only trends tool returning interest over time, interest by region, and rising or top related queries and topics for arbitrary search terms as structured JSON. Public GitHub repository created August 24, 2026.

Hosted Streamable HTTP MCP server from HasData for DuckDuckGo search results. It returns ranked organic results, ads in a separate array, DuckDuckGo AI answers and region-targeted search output as structured JSON for MCP clients. Public GitHub repository created August 20, 2026.

Stdio MCP server for evidence-grade capture of public URLs through the Attest Capture API. Its capture_url tool returns screenshot, PDF or HTML artifacts with a SHA-256 digest, timestamp and optional public evidence page for audit trails. Public GitHub repository created August 25, 2026.

Web search and fetch MCP server with built-in caching, rate limiting and content extraction. The current MCP surface exposes core search, fetch, crawl, extract and cache tools with zero API keys required, plus CLI and Python interfaces. Public GitHub repository created August 26, 2026.

Official Scrapy MCP server that connects MCP-capable agents to live Scrapy crawls for inspection, debugging, and steering. It talks over stdio to the agent and over HTTP with bearer auth to Scrapy RemoteControl job files, letting agents inspect crawl progress, settings, errors, active downloads, and candidate selectors. Public GitHub repository created August 27, 2026.

Official remote HTTP MCP server for GoAnyAPI, connecting agents to web data, Google Search, SEO, backlinks, Google Ads Transparency, traffic, domain intelligence, and related APIs. It uses OAuth 2.1 with PKCE so users authorize in the browser instead of copying API keys into local configs. Public GitHub repository created August 26, 2026.

Local SEO evidence and audit-automation toolkit with CLI, Docker, and stdio MCP surfaces. It analyzes Screaming Frog exports against a 96-check registry, exposes 42 shared SEO tools plus five Screaming Frog workflow tools, preserves skipped or unavailable evidence explicitly, and can generate structured audits, task backlogs, and report artifacts. Public GitHub repository created August 29, 2026.

Self-contained WordPress MCP server plugin that exposes WordPress through local WP-CLI stdio or remote Streamable HTTP. It maps WordPress Abilities API capabilities into 56 content, media, user, plugin, menu, widget, comment, options, and system abilities, with no cloud relay or companion plugin. Public GitHub repository created August 30, 2026.

Keyless-first TypeScript MCP server for web research, using DuckDuckGo by default and optional Brave or Tavily backends. It combines web search with a three-tier extraction ladder from fast HTTP fetch to hydration and Playwright, plus cache and SSRF protections, so agents get cleaner research payloads without mandatory API keys. Public GitHub repository created August 29, 2026.

Official 2Captcha MCP server for agents that need web search, web scraping, browser automation, marketplace parsing, structured extraction, batch jobs, and CAPTCHA solving against sites that block ordinary HTTP clients. It offers a hosted Streamable HTTP endpoint plus a local npx stdio server, both authenticated with a 2Captcha API token. Public GitHub repository created August 28, 2026.

Unofficial TypeScript MCP server for personal Ozon shopping research. It uses a headless Chrome session rather than Seller API keys to search products, retrieve prices, inspect product cards, compare listings, and read reviews through a local stdio MCP server intended for interactive use rather than bulk scraping. Public GitHub repository created September 4, 2026.

Composable TypeScript MCP server for PC game modding research and diagnostics. It bridges Nexus Mods, mod.io, Thunderstore, and a local workbench so agents can search mods, read changelogs, inspect versions, review dependencies, examine load orders, and trace crash logs without giving the assistant a broad browser or shell surface. Public GitHub repository created September 7, 2026.

Least-privilege JavaScript browser-reading MCP server for Savage Scraper. It opens explicitly allowlisted URLs in a dedicated tab of the user's existing Chrome profile, performs bounded lazy-load scrolling, and returns simplified page HTML through three stdio tools while intentionally omitting click, type, submit, arbitrary JavaScript, and unrestricted browser-control capabilities. Public GitHub repository created September 6, 2026.

MCP server and companion design skill from PyModel for grounding coding-agent UI work in real product screens. The hosted HTTP server or local npx stdio server exposes two tools for Niblet screen references and design-skill retrieval, with token-based access and production origins configurable through environment variables. Public GitHub repository created September 6, 2026 and updated September 8, 2026.

Standalone browser-content MCP server for JavaScript-rendered pages. It uses official MCP 2026-07-28 stdio transport and a server-owned Playwright Chromium profile to fetch bounded, inspectable page snapshots, apply extraction recipes, paginate cursor-protected results, and block unsafe URL forms by default. Public GitHub repository created September 8, 2026.

Multi-source ebook search tool delivered as a CLI, MCP server, and Agent Skill shell. It searches public-domain and open-license Chinese and English ebook sources, ranks links, and only downloads files when explicitly requested, using one core engine across all front ends. Public GitHub repository created September 9, 2026.

Local-first browser and web-reading MCP server for Claude Code, Claude Desktop, Cursor, and other clients. It offers budgeted page reads with explicit unread-region accounting, structured extraction with named evidence, real-browser lanes, monitors, workflows, and consent-gated capability packs, starting read-only so pages cannot trigger absent mutation tools. Public GitHub repository created September 9, 2026.

SEO and generative-engine-optimization MCP server with roughly eighty tools spanning Google Search Console, GA4, PageSpeed, CrUX, Knowledge Graph, Wikidata, IndexNow, WordPress over SSH, GitHub, structured data, llms.txt, AI crawler checks, E-E-A-T audits, and site migration checks. It supports stdio and stateless Streamable HTTP with read-only mode and toolset filtering. Public GitHub repository created September 9, 2026.

High-efficiency Chrome browser automation MCP platform that pairs an MV3 extension with a local native bridge so agents can control the user's active authenticated browser instead of a throwaway profile. It exposes roughly fifty-two tools, compact DOM and accessibility-tree views, network request helpers, chained JavaScript execution, visual fallback, tab lifecycle management, and a local HTTP MCP endpoint. Public GitHub repository created September 10, 2026.

Read-only Python MCP server for Google Search Console. It uses service-account authentication to expose site lists, search analytics, period comparisons, submitted sitemaps, and URL inspection results through the webmasters.readonly scope for Claude, Cursor, Claude Code, and other MCP clients. Public GitHub repository created September 14, 2026.

JavaScript MCP server for browser automation and API testing built on Playwright and Chrome DevTools Protocol. It gives agents numbered element references, Set-of-Mark screenshots, flow record and replay, element memory, captcha handoff, tab control, screenshots, and HTTP API test tools. Public GitHub repository created September 15, 2026.

JavaScript stdio MCP server for ClickTrail attribution implementation. It inspects projects, detects attribution gaps, plans installation, captures click ID schemas, validates pipelines, diagnoses missing identifiers, calculates coverage, reconciles conversions, and builds provider payloads. Public GitHub repository created September 15, 2026.

TypeScript MCP server for PromptEye brand-visibility analytics. It connects assistants to PromptEye projects, knowledge-base records, prompt suggestions, tracked prompt context, competitor visibility, answer/source views, and citation-quality reporting, with API-backed tools where live endpoints exist and explicit fixture fallbacks for not-yet-live metrics. Public GitHub repository created September 16, 2026.

Official TypeScript stdio MCP server for Glubler.com. It lets assistants search the Glubler catalog, inspect public Glubler lists, authenticate through a browser/token flow, create nested collections, glue single or batched websites into lists, and auto-fetch metadata and favicons for missing websites. Public GitHub repository created September 17, 2026.

Remote HTTP MCP server for Superflow, a visual collaboration and AI review layer for websites. It lets Claude, Cursor, and other MCP clients read website comments and AI review-agent findings, run review agents against URLs, inspect projects and pages, and authenticate either by bearer token or OAuth custom connector flow. Public GitHub repository created September 17, 2026.

JavaScript stdio MCP server from The MONA Group for searching, reserving, and buying .vn and international domains from Claude Code, Codex, Cursor, or Gemini. It is an alias of monacloud-mcp, exposes cloud_domain tools for pricing, reservation, VietQR payment, MONA Pass account claiming, and DNS attachment, and was published publicly on GitHub on September 19, 2026.

Python MCP server for searching Kleinanzeigen listings and fetching full ad, photo, location, attribute, and seller details. It wraps the Kleinanzeigen mobile APIs, adds server-side filters such as title-only, required terms, exclusions, price, category, radius, postcode, and seller type, and can download listing photos for vision-capable inspection. Public GitHub repository created September 13, 2026.

TypeScript MCP server for self-hosted SearXNG metasearch. It supports stdio by default and optional Streamable HTTP, exposes web, image, news, video, music, and page-fetch tools, requires no search API keys, and hardens page fetches with SSRF and DNS-rebind protections plus prompt-injection wrapping. Public GitHub repository created September 20, 2026.

Node stdio-to-Streamable-HTTP bridge for the hosted ViaFrei MCP service, which provides live German road, rail, parking, charging, fuel, weather-warning, and environmental-zone answers to MCP clients. The bridge is for stdio-only clients, holds no data or credentials, refuses cross-origin redirects, and relays to the public hosted endpoint with mandatory source attribution in results. Public GitHub repository created September 20, 2026.

Free open-source Node stdio MCP server that gives AI coding and design agents visual inspiration from Awwwards. It supports natural-language search over an FTS5 index, returns real screenshots inline, exposes site design DNA including palettes, technologies, elements, awards, and categories, and optionally uses Playwright and ffmpeg to capture fresh pages, analyze page structure, and record motion. Public GitHub repository created September 18, 2026.

Python MCP server for driving the real Chrome profile the user is already logged into, instead of launching a fresh automation browser. It connects over CDP or uses a macOS AppleScript fallback, provides token-saving page and DOM reads, real click and form interactions, screenshots, and background browser automation for Claude, Cursor, Antigravity, Cline, and other MCP clients. Public GitHub repository created September 17, 2026.

Rust SEO and generative-engine-optimization CLI with a stdio MCP server for local page audits, live crawls, structured-data checks, citation-readiness scoring, and CI gates. Core rule checks work without a model key; optional Jev features use TypeSafe API credits. Public repository created September 18, 2026.

Self-hosted search aggregator with local stdio MCP alongside HTTP API and CLI interfaces. It normalizes multiple search providers and manages keys, quotas, circuit breakers and failover; upstream provider credentials and local configuration are required for useful search. Public repository created September 21, 2026.

Node stdio MCP server and CLI for Semrush keyword, domain, backlink and competitor reports. Requires a Semrush API key and appropriate subscription or purchased API units; supported reports depend on the account's entitlement. Repository created September 25, 2026.

Chrome MV3 extension plus local stdio MCP server that lets an agent inspect and interact with the user's real browser tabs via a local broker. It uses logged-in browser context, so an MCP client should be trusted and browser actions supervised. Repository created September 20, 2026.

Unofficial read-only remote HTTP MCP server for Torob product searches, Iranian shop listings, variants and prices in Toman. No account or API key is required; prices and availability are read from Torob with a short in-memory cache, so confirm them before buying. Repository created September 24, 2026.

Self-hosted Vercel HTTP MCP connector for searching Reddit opinions and reading threads alongside web research. Requires approved Reddit API credentials and uses a secret embedded in the connector URL, which must be kept private. Read-only: no posting or voting. Repository created September 24, 2026.

Unofficial local stdio MCP server for the Czech Vilgain store: product search, ingredients, allergens, nutrition, cart edits and order history. Requires Vilgain account credentials; other regional sites are untested and checkout is intentionally left to the user. Repository created September 22, 2026.

Hosted Google Ads-specific HTTP MCP endpoint for campaign and keyword reports, search terms, budgets and ads. Markifact says reads run immediately while account mutations await approval; connecting a Google Ads manager account can expose multiple client accounts. Requires Markifact signup, Google Ads connection and OAuth; provider usage is credit-metered. Repository created September 28, 2026.

Unofficial local MCP server and CLI using an authenticated Uber web session to inspect fares, rides, trip receipts and some Eats data. Ride bookings and cancellations use preview and single-use confirmation tokens with fare ceilings and an audit log; this uses private web APIs that may change or conflict with platform terms. Repository created September 28, 2026.

CLI and stdio MCP for testing websites against plain-language goals and collecting verdicts from real pages. Supports browser-backed projects and recurring runs; its own documentation also describes native-app and game testing. Repository created October 1, 2026.

MCP bridge to existing Chrome or Brave profiles using a native-messaging extension, per-session stdio server and local broker. Provides profile-scoped tab control, audit log and human panic switch without a TCP listener; browser actions may change logged-in accounts. Repository created September 27, 2026.