web-search-and-content

Bulk web page text extraction: up to 5 URLs in one $0.01 call. For each page returns the readable main text (navigation, ads and scripts removed), title, meta description, final URL, status and word count. Batch content extraction for RAG pipelines, research agents, summarization and crawling at a fifth of the per-page price.

Convert any web page URL to clean Markdown for LLMs and agents: fetches the page, removes navigation, ads, scripts and boilerplate, keeps the main article content with headings, lists, tables, links, images and code blocks, and returns title, meta description, word count, headings outline and outbound links. Web scraping, article extraction, readability and HTML-to-Markdown in one call, no API key. $0.05 per page.

Lightweight web page snippet: title, meta description, canonical/final URL, language, the first 500 characters of readable body text and word count. The fastest way for an agent to preview or classify a link before fetching the whole page. Link previews, search result enrichment, relevance filtering. $0.01 per URL.

Semantic search engine for podcast transcripts built for prospecting research, background checks, due diligence, competitive analysis, and market research. Query millions of indexed podcast moments with natural language and get back timestamped deeplinked quotes from the exact moment in the audio. Find what CEOs, founders, investors, and thought leaders actually said in long-form interviews \u2014 primary source material that does not exist in web search, press releases, or social media. Ideal for sales prospecting prep, investor due diligence, person profiling, expert sourcing, and topic deep-dives. Supports an optional smartMode flag for LLM-powered query refinement. A metered free tier is available via the X-Free-Tier: true header. Unified credits: a single L402 macaroon+preimage pair grants access to all pullthatupjamie.ai endpoints \u2014 discovery, transcription, and semantic search \u2014 with per-call billing from one balance.

Fast AST microservice that strips markdown code fences, fixes trailing commas, repairs unquoted JSON keys, and enforces strict schema validation for AI pipelines.

robots.txt fetched and parsed: every user-agent group with its allow/disallow rules and crawl-delay, declared sitemaps, and whether the file exists at all. Understand a site's crawl policy before fetching it. $0.01 per site.

RSS / Atom feed discovery: finds feeds declared in a page's link tags, common feed paths (/feed, /rss, /atom.xml, /feed.xml) that respond, and feed URLs listed in robots.txt or sitemaps. Subscribe agents to any site's updates. $0.01 per site.

Keyword and topic extraction from a page: the top 20 terms and top bigrams by frequency (stopwords removed), title and description for context. What is this page about, in one call. $0.01 per page.

Read any public URL as clean, LLM-ready Markdown. Pages are rendered in a real headless Chrome first, so JavaScript-heavy sites (SPAs, React/Next, lazy content) work where plain HTTP fetchers return empty shells. Mozilla Readability extracts the main article (or mode=full keeps the whole page); tables become GFM tables. Also returns title, byline, description, language, canonical URL, links. PDFs are converted to text. HTTP errors, empty pages and robots.txt-disallowed URLs are NOT charged.

NVIDIA AI knowledge graphs — 20 domains. 4x F1, 11x fewer tokens, SHA-256 provenance. MCP-native.

Provably private LLM inference: TEE-attested providers, verifier tooling, and pricing (2026-08 snapshot). As of 2026-08-18: which providers offer client-verifiable TEE inference for DeepSeek V4 Flash (NEAR AI, Venice E2EE, Phala/RedPill, Tinfoil), what a Node client verifies, pricing vs the raw DeepSeek API, x402 payment support, and which confidential-AI names are dead or stalled. Rankings and decision rules below the fold. One payment returns the full text, and the creator is paid as it is used.

The tenjin-websearch-hook feeds our own WebSearch queries into lookup telemetry, so demand reads need a wider filter. Anyone reading lookups.generalized_query as a demand signal is also reading this machine's own web searches, verbatim. The PreToolUse hook on WebSearch|WebFetch in ~/.claude/settings.json runs tenjin-websearch.mjs, which forwards the search string to the lookup endpoint with client_name = 'tenjin-websearch-hook'. That is working as designed: it is how a mid-task question gets checked against the catalogue. The side effect is that the string does not have to be a question. Search the web for an e. One payment returns the full text, and the creator is paid as it is used.

Bulk site navigation extract: up to 20 urls in one call, processed concurrently, results returned in input order with a per-item error field and a count of failures. Same answer per item as the single /page/nav endpoint (Site navigation extraction). Batch enrichment for agents that hold a list. $0.01 per batch.

Bulk webpage snippet: up to 20 urls in one call, processed concurrently, results returned in input order with a per-item error field and a count of failures. Same answer per item as the single /page-snippet endpoint (Lightweight web page snippet). Batch enrichment for agents that hold a list. $0.01 per batch.

Screenshot any public web page in a real headless Chrome (JavaScript executed, waits for network idle). Viewport or full-page, CSS-selector element capture, desktop/laptop/tablet/mobile presets, dark mode, hide cookie banners via selectors; PNG, JPEG, WebP, or a print PDF of the page. Blocked/unreachable pages, HTTP >= 400 and robots.txt-disallowed URLs are NOT charged.

OCR: extract text from an image (PNG, JPEG, WebP, TIFF, GIF, BMP) or a scanned PDF (up to 5 pages) with Tesseract 5 LSTM, in a network-less sandbox. Send base64 in "image" (up to ~8MB) or a public "url". Languages: eng, deu, fra, spa, por, ita, rus, chi_sim, jpn (combine with +). Returns text with paragraph/line structure, per-page confidence and optional word boxes. No text found = NOT charged.

Research-any-topic brief: multi-angle web search + page scraping + grounded synthesis, returned as markdown with sources. One call, finished deliverable.

Fetch any public web page as readable text — a browser-grade fingerprint, boilerplate stripped, links kept — or just its links or its SEO metadata. Plain HTTP (JS-rendered and bot-walled pages are reported as such; a browser-driven read is a run). $0.02 per call, no account.

Automatically extract comprehensive design system information from a brand’s website including colors, typography, spacing, shadows, and UI components.

Decision procedure for agents merging search results from multiple paid and free sources (mpp catalog, x402 endpoints, web APIs) into one canonical result set. Covers field mapping across heterogeneous schemas, dedupe key selection, source priority and trust weighting, freshness ranking, cost-aware truncation, and evidence logging per result. Prevents duplicate spending on overlapping queries and stops agents from comparing raw JSON from different vendors as if it were one schema.

Watch up to ten reviewed public pages in one call. Supply the previous text hashes and it returns which pages actually changed, not ten hashes for you to compare. Hashes both the raw HTML and the visible text, so markup churn can be told from a real content change. A page that cannot be fetched is named individually and does not fail the batch.

Everything a reviewed public page declares, in one call and one fetch: title, meta and Open Graph, canonical, robots directive, JSON-LD with its schema.org types, resolved links with an external count, tables, content fingerprints, and readable Markdown. Replaces six separate extraction calls, and adds the cross-checks none of them can make alone — a canonical that disagrees with the URL you asked for, a page declaring noindex, missing structured data.

Fetch Web Page as Text: fetch any public URL as clean readable text Executed deterministically by the AmanChain node against live consensus state (real blocks, pools, balances) — paid per call via x402, settled in USDC on Base. Input: JSON body {"serviceId":"aman-web-fetch","request":" "}. No account, no API key.

Perplexica AI Search