web-search-and-content

Sovereign, local-first Jamaican Patois Translation API with Lightning paywalls. Translates standard English into fluent, authentic Jamaican Patois (Patwa) with natural spelling, phonetics, and customizable vibes (music, farming, weather, default). Powered by your home-hosted RTX 3060.

Perplexica AI Search

实时网页搜索 — 实时网页搜索：输入关键词，返回最新搜索结果(标题+链接+摘要)，免费源无需API key。适合 agent 实时信息检索、新闻追踪、事实核查、研究资料收集——生态最值钱的能力之一，补上我们最后的短板。

Audio-to-text conversion: submit any public audio file URL (mp3, wav, m4a, ogg, up to 25MB) and get back a full text transcript, with the source language automatically detected across roughly 99 languages. Audio is processed in an isolated sandbox and deleted immediately afterward. Built for turning meeting recordings, interviews, and voice memos into searchable text.

Speech-to-text API: audio URL in, transcript out — a neural speech-recognition engine on our own hardware, ~99 languages, mp3/wav/m4a/ogg up to 25MB. For voicemail, podcast, meeting, and voice-agent pipelines. Audio deleted after processing.

Video-to-audio extraction: pulls the audio track out of any public video or audio URL (up to 25MB) and returns it as an MP3, encoded on request rather than pulled from a pre-made library. Source files are deleted the moment processing finishes. Use it to prep clips for transcription, strip audio from screen recordings, or normalize mixed formats before a media pipeline.

reCAPTCHA v2/v3, hCaptcha, Turnstile. 🆓 Free trial — no signup, no human. ?trial=1

Text statistics API: characters, words, sentences, average sentence length, Flesch reading-ease score, estimated reading time, and estimated LLM token count for any text — instant, computed locally, nothing stored. For context-budget checks, readability gates, and content QA in agent pipelines.

URL unshortener API: follow any short link safely and get the final destination, full redirect chain, and final status — without visiting it yourself. Link-safety pre-check for agents clicking the unknown web.

Tech news front page API: the current top 30 ranked tech stories — title, outbound URL, points, comment count, and timestamps — as normalized JSON. For trend detection, topic monitoring, and content-pipeline agents. Refreshed every few minutes.

Page metadata API: extract the title, canonical URL, favicon, OpenGraph tags, Twitter card tags, description, author, and keywords from any public web page as clean JSON — the everything-you-need for link previews, SEO audits, share cards, and content catalogs. SSRF-guarded, 512KB cap.

PDF to text API: fetches any public PDF URL and returns its full plain text, fast, with layout preserved. For scanned PDFs needing OCR see /pdf-to-markdown. 25MB / ~100 page cap. Nothing stored.

PDF text extraction API: give it any public PDF URL and get clean layout-preserving text back in one call. Handles digital PDFs up to 25MB. Use it as a PDF parser, PDF-to-text converter, or document ingestion step for RAG and analysis agents. Files are processed in a temp sandbox and deleted immediately.

Extracts clean, layout-preserving text from any public PDF URL up to 25MB in a single call, no local PDF library or OCR pipeline required. Built for feeding PDF content into research agents, document QA, and RAG ingestion pipelines that need plain text rather than a binary file.

HTML to Markdown API: pass a public URL (or raw HTML) and get clean readable markdown back — title, byline, article body with navigation/ads/boilerplate stripped. We honor robots.txt AND Content-Signal (aipref) declarations: a site that disallows agent access returns an unpaid 403, you are never charged. Use for research agents, summarization pipelines, and RAG ingestion.

PDF parser API for agents: URL in, extracted text out. Digital PDFs parsed with layout preserved. Use for document ingestion, RAG pipelines, contract reading, and report analysis. No retention.

Language detection API: identify the language of any text (60+ languages) with confidence scores and top-3 candidates — computed locally in milliseconds, text never stored. For routing multilingual content, translation pre-checks, and moderation pipelines.

Fetches a domain's robots.txt and parses its Content-Signal and AI-preference directives (search, ai-input, ai-train) into structured JSON plus a plain-English summary of what's allowed — the "can this domain be crawled by an AI agent" check. Run before scraping content or building a RAG index, ahead of the industry's move toward stricter bot-gating defaults later this year.

Webpage-to-markdown extraction: submit any public URL and receive the article body as clean markdown, with navigation, ads, and boilerplate stripped and the byline preserved. Respects each site's crawl permissions automatically. Built for feeding readable page content into summarizers, search indexes, and language-model context windows.

Readable-article extraction API: strips a public web page down to its title, author line, and main body text, discarding menus, ads, and sidebars. A pre-crawl permission check runs first, honoring each site's crawl rules. Useful for building article archives, feeding LLM pipelines, or preparing pages for text-to-speech and translation.

Extracts structured data from content using AI processing via a relay interface.

Slang lookup and browse API: query a specific term or sample random entries from one generation's vocabulary — each entry carries a one-sentence meaning, its origin generation, approximate era, and an example sentence. Local curated dataset, same answer every time for the same term.

Full-text search across the entire indexed public-domain corpus. Returns the most relevant passages from any book with title/author citations, ranked by relevance. The premium cross-corpus retrieval route — no LLM generation, real excerpts with sources. No license, no API key required.

Web Reader ($0.003/call): Fetch any public web page and return its main content as clean, boilerplate-free markdown (reader mode for agents). Handles redirects, strips nav/ads/scripts, reports title, byline, language and word count.