InfinityScrape MCP
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PYTHONUNBUFFERED | No | Set to 1 to force Python stdout/stderr to be unbuffered, ensuring stdio JSON-RPC messages are flushed immediately when running the MCP server via `python -m infinity_scraper.server`. | 1 |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| scrape_urlC | Scrape any website and convert its content into clean, ad-free Markdown. |
| search_webC | Perform a live web search to find relevant URLs, articles, and snippets for any topic. |
| search_and_scrapeB | Autonomous research tool: searches the web and automatically scrapes the top results into a synthesized Markdown report. |
| scrape_dynamicB | Scrape dynamic Single Page Applications (React, Vue, Next.js) with Playwright headless browser. |
| extract_schemaB | Extract specific structured fields from a webpage using a key-to-CSS-selector dictionary map. |
| deep_crawlC | Recursively crawl a website or documentation tree up to max_depth and aggregate into unified documentation. |
| extract_structuredC | Extract structured data (JSON-LD schema, HTML tables, metadata, and CSS-targeted nodes) as JSON. |
| scrape_batchC | Concurrently scrape multiple URLs in parallel and return their combined Markdown. |
| extract_pdfB | Download and extract full text and metadata from an online PDF document. |
| get_youtube_transcriptC | Extract full text transcript and timestamps from any YouTube video URL in milliseconds (Zero GPU). |
| osint_deep_public_reconC | Ultra-Precision Public Intelligence Dossier Tool: Scrapes and correlates open web data across 25+ platforms with negative exclusions, PIN/Street drill-down, and multi-signal Confidence Scoring (0% - 100%). |
| osint_geoint_lookupC | GEOINT Location Tool: Resolve any address, landmark, or street into a complete global administrative hierarchy. |
| osint_location_entity_searchC | Hierarchical Location-based OSINT Search Matrix with Negative Exclusion Filters. |
| osint_domain_reconC | OSINT Tool: Gather public domain intelligence including SSL Certificate validity/issuer, DNS resolution, and WHOIS/RDAP. |
| osint_tech_stackC | OSINT Tool: Fingerprint web technology stack, frameworks (React, Vue, Next.js, Angular, Django, etc.), CMS, CDN, and servers. |
| osint_ip_lookupB | OSINT Tool: Lookup public IP Geolocation, ASN, ISP, and Organization information. |
| osint_username_checkB | OSINT Tool: Check public presence and profile existence for a handle/username across 25+ developer and creator platforms. |
| osint_searchC | OSINT Tool: Targeted intelligence search / dorking using advanced boolean operators (site:, filetype:pdf/xlsx/docx, intitle:, -exclude). |
| osint_wayback_time_machineA | Wayback Machine Historical Recon Tool: Time-travel and retrieve archived snapshots of deleted or past webpages from the Internet Archive. Args: url: The webpage URL to investigate. timestamp: Optional 14-digit timestamp (e.g. '20210515120000'). If empty, fetches most recent. list_snapshots: If True, returns a list of available historical snapshot dates instead of page content. limit: Number of snapshot dates to return (default 5). |
| osint_subdomain_enumerationB | Subdomain Recon Tool: Enumerate all public active and historical subdomains for any domain via Certificate Transparency Logs in <1 second. Args: domain: The root domain to scan (e.g. 'target.com' or 'github.com'). limit: Maximum subdomains to return (default 50). |
| osint_dns_auditC | DNS & Infrastructure Audit Tool: Lookup IPv4, IPv6, and server infrastructure records for any domain. Args: domain: Domain name to audit (e.g. 'openai.com'). |
| extract_image_exifA | Image EXIF & Geotag Recon Tool: Extracts camera make/model, creation timestamp, editing software, and exact GPS coordinates (with map link) from any online image. Args: image_url: The direct HTTP/HTTPS URL of the image file (.jpg, .png, .tiff, .webp). |
| extract_reddit_thread_toolA | Reddit Discussion Thread Extractor: Extracts full Reddit post content, upvote ratio, author, and top community comments into structured dialogue Markdown. Args: url: Full Reddit post URL (e.g. 'https://www.reddit.com/r/programming/comments/...'). max_comments: Number of top comments to extract (default 25). |
| extract_rss_feed_toolA | RSS & Atom Feed Ingestion Tool: Extracts recent articles, publish dates, links, and summaries from any RSS or Atom feed (Substack, blogs, news). Args: feed_url: The RSS or Atom feed URL (e.g. 'https://openai.com/news/rss.xml' or 'https://blog.bytebytego.com/feed'). max_items: Number of recent articles to extract (default 10). |
| optimize_rag_chunksA | Semantic RAG Chunker & Token Optimizer: Splits large scraped web documents into header-aware semantic chunks with token counts and section headers for LLM embeddings and context efficiency. Args: text_or_markdown: The large Markdown or text string to chunk. max_chunk_chars: Maximum characters per chunk (default 2000 ~500 tokens). |
| extract_contactsA | Zero-Selector Contact & Lead Intelligence: Heuristically extracts emails, phone numbers, WhatsApp links, social profiles, physical addresses, PIN/ZIP codes, and business registration IDs (GSTIN/CIN/EIN) from a webpage or raw text. Args: url: The webpage URL to scrape and analyze (optional if html_text is given). html_text: Raw HTML or text to analyze directly (optional if url is given). |
| extract_pricing_tableB | Zero-Selector SaaS & Product Pricing Table Extractor: Automatically identifies tier names (Free, Pro, Enterprise), price amounts ($/mo, $/yr, INR), billing cycles, feature bullet lists, popular plan badges, and CTA links. Args: url: The pricing page URL to scrape and analyze. html_text: Raw HTML or text of the pricing section. |
| extract_faqsB | Zero-Selector FAQ & Accordion Intelligence Extractor: Extracts structured question-and-answer pairs by inspecting JSON-LD FAQPage schemas, HTML5 tags, and accordion CSS heuristics. Args: url: The webpage URL to extract FAQs from. html_text: Raw HTML to extract FAQs from. |
| crawl_sitemapB | High-Speed XML Sitemap & Index Parser: Streams and parses XML sitemaps, handling namespaces, .xml.gz gzip streams, and automatically crawls child sitemaps if a sitemap index is provided. Args: sitemap_url: URL of the sitemap (e.g. 'https://example.com/sitemap.xml'). max_urls: Maximum URLs to extract (default 200). crawl_nested: If true, automatically follows nested sitemaps inside sitemapindex. |
| discover_sitemapsA | Autonomously discovers all active XML sitemaps for a website by reading robots.txt and testing standard location heuristics (/sitemap.xml, /sitemap_index.xml, etc.). Args: base_url: The root domain or homepage URL (e.g. 'https://stripe.com'). |
| extract_docx_documentA | Word Document (.docx) Streamer & Table Parser: Streams and converts remote or local Word .docx files into clean Markdown preserving headings (H1-H4), bullet points, and full table matrices without local disk bloat. Args: url_or_path: Remote URL (http/https) or local absolute file path to the .docx file. |
| extract_xlsx_spreadsheetA | Excel Spreadsheet (.xlsx) Streamer & Structured Table Extractor: Streams remote or local .xlsx spreadsheets and parses all worksheets into clean Markdown tables and structured JSON records. Args: url_or_path: Remote URL (http/https) or local absolute file path to the .xlsx file. max_rows: Maximum rows to extract per worksheet (default 100). |
| compare_page_diffA | Temporal Page Diff & Change Radar: Compares two versions of a webpage or text, detecting pricing shifts, added/removed feature sections, and calculating overall content similarity percentage. Args: old_content_or_url: Historical text/markdown or URL. new_content_or_url: Current live text/markdown or URL. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 33 tools
Several clusters of tools have unclear boundaries: scrape_url vs scrape_dynamic vs scrape_batch vs deep_crawl vs search_and_scrape all fetch page content, and extract_schema vs extract_structured both do field-level extraction. The OSINT cluster (osint_domain_recon, osint_dns_audit, osint_subdomain_enumeration, osint_tech_stack, osint_ip_lookup) and osint_geoint_lookup vs osint_location_entity_search overlap heavily, forcing guesswork about which to pick.
Most tools follow snake_case verb_noun (scrape_url, extract_pdf, search_web), but there are deviations: two tools carry a redundant '_tool' suffix (extract_reddit_thread_tool, extract_rss_feed_tool), some use a domain prefix (osint_*), and others are bare noun_verb or non-verb forms (deep_crawl, scrape_batch). Readable overall but inconsistent conventions.
33 tools is well above the comfortable range for a single server and includes near-duplicate scraping and OSINT entries that could be merged. The breadth (scraping, document parsing, OSINT, RAG chunking) makes the surface feel like several servers bundled together.
The scraping/extraction/OSINT domain is well covered: static and dynamic scraping, crawling, sitemaps, PDF/DOCX/XLSX, images, YouTube, Reddit, RSS, and a broad OSINT suite, plus RAG chunking and diffing. Minor gaps like screenshot capture, authenticated/form-based scraping, and a generic 'get robots.txt' helper are workable around.