information-retrieval

Get a recap of the news or summary of a covered news topic over the past 12-24 hours.

Harvests the curated feed set for one topic and language, groups the headlines into events and returns one entry per event. The source count is the product: ten outlets on one story is a different signal than one. Never returns third party article bodies, only headline, own abstract and link.

List Reels from a Facebook profile or Page.

List comments on a Reddit post by URL.

Search Reddit posts site-wide by keyword (not scoped to one subreddit — use reddit.subreddit.search.list for that).

List comments on an Instagram post or Reel by URL.

Get the transcript for a Facebook post or Reel by URL.

Get structured metadata for a public webpage: title, description, canonical URL, Open Graph fields, favicon, and language.

DNS propagation checker. Compares a DNS record across public resolvers (Google, Cloudflare, Quad9, OpenDNS) to spot propagation lag or split-horizon answers.

Encyclopedic grounding in one call: Wikipedia search plus the full summary of the best match — extract, short description, canonical URL, thumbnail — in any of 300+ language editions. Official Wikimedia REST APIs, CC BY-SA attribution included. The cheapest way for an agent to ground an entity before reasoning about it.

Classify text into your labels with confidence and per-label scores. Single- or multi-label. Cheap and high-volume: routing, tagging, moderation triage, intent.

Search current Google News results. Returns headline, source/domain, URL, snippet, publication time, image and rank. For current-events, monitoring and research agents.

Fetch a URL and report whether it is reachable, the HTTP status code, and the response time in milliseconds. A fast liveness/uptime probe — run it before relying on a site or as a cheap health signal.

Live YouTube video-search results for a query, localised by location and language. Returns ranked videos with title, URL, video id, channel, view count, publish date and duration.

News search: live Google News results for a query, localised by location and language. Returns ranked articles with title, URL, domain and, where available, source and timestamp. Current events and headlines for agents.

UUIDs, hashes, base64, URL encoding, secure passwords and timestamp conversion. Deterministic, no inference, answers immediately. POST {"op":"uuid","n":2}. Use it when your agent needs a fresh UUID, hash, password or encoding and cannot run code itself.

Wikidata firmographics (keyless, CC0)

Fetch a URL and extract every hyperlink on the page, each resolved to an absolute URL and classified as internal vs external (by registrable host), with its anchor text, rel attributes (nofollow, sponsored, ugc) and whether it opens in a new tab. Returns internal/external/nofollow counts. For SEO link audits, internal-linking analysis and outbound-link/nofollow compliance checks.

Read a single XBRL financial concept for a company (by CIK or ticker) from its as-filed statements. Give the concept's XBRL tag (e.g. Revenues, NetIncomeLoss, Assets, EarningsPerShareDiluted) and optionally the taxonomy (default us-gaap, also dei). Returns the label and, per unit (USD, shares), recent datapoints with period end, value, fiscal year/period, form, filing date and accession. Bounded, as-reported output.

Fetch a public page and grade its on-page SEO: title and meta-description length, H1 and heading outline, canonical, meta-robots indexability, viewport/lang/charset, hreflang, Open Graph and Twitter Card completeness, image alt coverage and word count — plus optional focus-keyword placement and density. Returns a pass/warn/fail verdict per check and an overall 0–100 score.

Fetch a site's robots.txt and evaluate crawl permissions: given a URL (plus optional extra paths) and a user-agent, return whether each path is allowed or disallowed, the rule that matched, the user-agent group, the crawl-delay and any declared sitemaps. Implements the Robots Exclusion Protocol (RFC 9309) with longest-match-wins, Allow-over-Disallow tie-breaking and * / $ wildcards.

Turn messy text into clean JSON. Pass text plus a field `schema` or plain `instructions`; get a validated JSON object back. Uses the provider JSON mode.

Research brief with citations: a quick, sourced answer to any question or topic. Use before writing, deciding, or searching deeper. Pass q (question). Returns a summary, 3-6 cited bullets and sources (official docs, Wikipedia, Stack Overflow, GitHub, papers) with URLs and dates; off-topic sources dropped. 0.01 USDC (Base/Polygon/Arbitrum/Solana/Avalanche/Sei). Errors are never charged: 400 bad input, 422 nothing found or target unreachable, 503 busy.

Search the periodic table by name or symbol substring. Exact matches rank first, then prefix matches, then substring matches. Returns up to `limit` elements (default 20) plus the total match count.