Find the obligatory xkcd for whatever you're working on. Article, slide deck, memo, PDF, blog post. Plug it into your AI tool. Your model searches our index of all xkcd comics annotated with applicability, humor type, and themes, picks the right one, and drops it inline.

Fleet entry point for Moltline Studio: 22 hosted MCP servers, 160 tools, 110 of them free. Searches 138 SKILL.md agent skills and routes you to the right server. No account, no API key, no signup on the free tier. Independently audited — MCPize Verified A. Direct endpoint: https://mcp.moltlinestudio.com/catalog

Geonode Scraper MCP connects AI assistants to web content over Streamable HTTP. Use the hosted extract tool without an account or API key to retrieve public pages as Markdown or HTML. Anonymous access supports synchronous extraction only: 5 requests per minute, one concurrent request, and 100 MB of downloaded content per day per network address. Shared networks may share allowances. Anonymous usage is metered but not billed. JavaScript rendering, country selection, and custom headers other than Accept-Language require an API key. For additional tools, including map, search, crawl, batch and job management, add your Geonode API key in the X-Api-Key header on the same endpoint. Authenticated usage follows your plan. Direct endpoint: https://scraper.geonode.io/mcp Setup: https://docs.geonode.com/docs/scraper-api/guides/mcp/01_use-without-an-api-key

Twitter/X scraping API for AI agents. Scrape public profiles (bio, stats, verification), user tweets (text, engagement, media), and search results — all without needing a Twitter API key. Structured JSON output ready for analysis. Tools: twitterscrapeprofile, twittersearchtweets, twittergetusertweets. Use this for social media monitoring, influencer research, sentiment analysis, competitive intelligence, or building social dashboards. The social intelligence layer for AI agents. IMPORTANT: For other social networks, use sociallookup_profile. Returns structured JSON with profiles, tweets, engagement metrics. No API key required — x402 micropayment $0.005/call on Base L2.

Post, schedule, queue and thread social media posts to X (Twitter) from your agent, with a human approval queue (needsapproval) so a person reviews a post before it goes out. Use when the user says "post this to X", "tweet this", "cross-post this", "share this launch", "announce the release", "schedule a tweet for tomorrow at 9", "queue these posts", "post this as a thread", "put this up for approval", "what's in my post queue", "reschedule that post" or "post a changelog update". 16 tools. Networks today: X is live for everyone; Threads after Meta app approval; Bluesky, LinkedIn and Mastodon in testing. No DMs, mention replies, video networks, analytics or team seats. Auth: API key as Authorization: Bearer op.... $29/$79/$199 a month, 7-day trial for $1.

Agent-first browser server: one Rust binary with V8 built in. Tiered fetch, multi-engine search, screenshots, downloads, click/eval, persistent login sessions, and render_markdown (markdown + archify diagram fences to deterministic HTML+SVG artifacts). Stealth TLS fingerprints, optional 2captcha integration, no Chromium.

Web search and page reading for AI agents. Search the web, open any page or JSON API, read RSS feeds and YouTube transcripts, and get the full text back with a source link for every result. Search returns ranked results with titles, links and snippets, ready to cite. Fetch opens any URL your agent builds — web pages, JSON REST APIs, RSS and Atom feeds — and returns the full content: pages as clean markdown, JSON and XML as the raw, parseable body. Pages that need JavaScript, refuse plain requests or rate-limit are opened in a full browser, and Levirge remembers which sources need it so the next visit is fast. Keep context small. select: return only the fields or sections you need schema: discover a response's shape first maxbytes: cap the size YouTube: a video link returns the title, channel, description and full timestamped transcript. Research at scale: queue up to 200 URLs with fetchasync and collect them later with fetch_results. Requests

Fast deterministic web utilities for autonomous agents: web metadata extraction and website change detection. Web Metadata extracts Open Graph, JSON-LD, canonical URLs, robots directives, headings, language, and page metadata. Web Change provides baseline capture, changed/unchanged detection, hashes, and deterministic diffs. Both are designed for recurring agent workflows such as SEO crawlers, metadata ingestion, price monitoring, documentation monitoring, competitor tracking, and RAG preprocessing.

vastlint - IAB XML VAST validator and linter

VAST XML validation MCP server for programmatic video advertising. Over $30 billion in annual CTV and video ad spend flows through VAST XML — malformed tags cause lost impressions, broken tracking, and revenue discrepancies between platforms. Validates VAST tags against IAB Tech Lab VAST 2.0, 3.0, 4.0, 4.1, 4.2, and 4.3 specifications, W3C XML 1.0 well-formedness, RFC 3986 URI syntax, IANA media types, ISO 4217 currency codes, Ad-ID registry format, and IAB SIMID 1.0–1.2. 118 rules covering required fields, schema validation, structural correctness, security, deprecated features, and value formats. Tools: validatevast (validate inline XML), validatevasturl (fetch and validate a VAST URL, following wrapper chains), listrules (browse all 118 rules), explainrule (get fix guidance for a specific rule), fixvast (auto-fix common issues). Built on a zero-dependency Rust core. Sub-millisecond latency. No data retention — VAST XML is validated ephemerally and never stored. Self-hostabl

X (Twitter) profiles and tweets by handle, URL or user ID — bio, follower counts, tweet engagement (likes, retweets, replies, views), single-tweet lookup by URL, engagement/text filters, incremental runs for scheduled monitoring. No login, no cookies, no X API key. Runs on Apify; you pay per result ($0.005/account, $0.0003/tweet, $0.005/tweet lookup) at the same price on every plan, and only for rows that actually came back. Bring your Apify token (free account works). Cannot do: keyword search, follower lists, reply threads — X does not expose these to logged-out clients and this tool does not use logged-in accounts.

Detect what changed on a page, a sitemap or a feed since the last check. Returns added, removed and modified items with a text diff and a stable change id. Can email a summary when something changes, and stays silent when nothing did. Works statelessly for agents or on a schedule.

Fetch web pages, PDFs and search results as compact, LLM-ready Markdown or ready-made chunks for vector databases (RAG). Tools: scrape a page, search the web, map a site's URLs, crawl several pages, extract structured JSON, split pages into chunks with heading paths, and scheduled page monitoring (beta). Firecrawl-compatible API, so the same account also works with the official Firecrawl SDKs. Sign-in by email code, no password; 1,000 free credits a month.

Stealth web scraping API for AI agents. Returns clean, LLM-ready Markdown from any URL with anti-bot bypass (Cloudflare Turnstile, Datadome). Native MCP integration, x402 micropayments ($0.005/request), multiple output formats (Markdown, screenshot, PDF, CSV, HTML). No accounts needed — agents pay per request with USDC on Base.

Web data API that just works. Pass any URL, get clean markdown — Cloudflare, JavaScript, CAPTCHAs handled automatically. 97.6% success rate across 500+ domains. 55+ built-in extractors for Amazon, Reddit, YouTube, Wikipedia, LinkedIn. 65-98% token savings. Zero config: npx webpeel mcp. Free tier: 2,000 fetches/month.

Find viral posts, outliers, hooks and ads on Instagram, TikTok, YouTube and Facebook, then write.

Schedule posts to 9 social networks from Claude, ChatGPT or Cursor.

Live Elon Musk & Ted Cruz tweet feed, real-time tweet counters, and market data for Polymarket's tweet-count markets. An independent counter polls X every ~60s, classifies posts against the actual settlement rules (replies mostly don't count, deleted posts do), and publishes three counters side by side — the real count, the XTracker oracle count, the UI count — plus the deltas between them. Free tier is keyless: getmarketmetadata, getrecenttweets, getpredictions, checkhealth work with no API key. A PRO key (x-api-key header) unlocks gettradingsignals, analyzeportfolio, and getrebalancing_recommendations. Docs: https://polystrike.xyz/docs/guides/mcp-setup/ · OpenAPI: https://polystrike.xyz/openapi.json · Examples (MIT): https://github.com/dizpers/polystrike-examples

Search public RSS, Atom, and RDF feed items by keyword, regex, or field, with exclusions. Check search status and retrieve matching items.

Lint a SKILL.md for frontmatter, structure, secrets and size. All 6 tools free. Part of Moltline Studio: 22 hosted MCP servers, 160 tools, 110 of them free — no account, no API key, no signup. Independently audited — MCPize Verified A. Direct endpoint: https://mcp.moltlinestudio.com/skillmd-lint

Crawl and scrape websites to clean markdown, and track what changed between crawls. Scrape one URL or a batch of up to 500, no setup needed Crawl a whole site once, or map every URL it declares in its sitemaps Watch sites over time as projects, and see exactly which pages were added, modified or removed Search inside stored pages, and read any single page in full Sign in with your MeshArc account. A read-only connection can read everything your workspace has already stored; tick write access to fetch new pages (this uses credits).

Adversarial test suites for AI-built code. Point your agent at a file; it returns a runnable suite, a coverage report, and the bugs the suite will expose - with code evidence.