Sift Web Tools

Gas and priority fees on 5 chains (slow, standard, fast), live. Current gas price, next base fee and priority-fee percentiles (slow, standard, fast) for ethereum, base, arbitrum, optimism and polygon, from the last 5 blocks. Pass chain=all (default) or one chain.

ENS lookup: .eth name to address, or address to verified name. Resolves an .eth name to its address, or an address to its primary ENS name (only returned when the forward record matches).

Find a site's RSS/Atom feeds and return the latest items. Discovers RSS or Atom feeds for a site or page and returns the latest items with title, link, date and summary.

ERC-20 token details: name, symbol, decimals, total supply. Reads an ERC-20 token contract on ethereum, base, arbitrum, optimism or polygon and returns name, symbol, decimals, total supply and bytecode size.

Transaction details: status, from, to, value, fee, token transfers. Looks up a transaction hash on ethereum, base, arbitrum, optimism or polygon: status, block, time, from, to, value, gas fee, method id and every ERC-20 transfer in it.

Wallet balance: native coin plus USDC/USDT/DAI, ENS name, on 5 chains. Balance of any wallet (0x address or .eth name) on ethereum, base, arbitrum, optimism or polygon: native coin, USDC, USDT and DAI, whether it is a contract, and its verified ENS name.

Page metadata: title, description, Open Graph, Twitter card, JSON-LD. All metadata of a web page: title, description, language, canonical URL, favicon, Open Graph, Twitter card and JSON-LD structured data.

Email address check: syntax, MX, disposable, role and free-provider flags. Validate an email address without sending mail: syntax, MX records, disposable-domain list, role account and free-provider flags, plus a verdict.

Extract all links from a page, split internal and external. Every link on a page with anchor text, deduplicated and split into internal and external.

Domain WHOIS via RDAP: registrar, creation and expiry dates, age. Registration data from the official RDAP registry: registrar, status, created, updated, expiry, domain age in days, nameservers, DNSSEC.

Token safety check: risk score, liquidity, supply held by owner and pools, upgrade, mint and blacklist powers. Check a token before trading it (Base, Ethereum, Arbitrum, Optimism, Polygon). Reads the contract from the chain: proxy, mint, blacklist, pause, fee and trading switches, owner. On Base and Ethereum also finds Uniswap v3 and Aerodrome pools, estimates liquidity in dollars and shows supply

Ask a question about a web page and get a grounded answer with quotes. Answers a question using only the content of a public page, says whether the answer was found, and returns verbatim supporting quotes.

Screenshot of a web page as PNG (headless Chromium). Renders a public page in headless Chromium and returns a PNG. Viewport width 320-1920, height 320-1600, optional full page.

Fetch a web page as clean Markdown. Readable main text of any public web page as Markdown, with title, author, date and word count. Boilerplate, menus and ads removed.

List a website's URLs from its sitemaps. Reads robots.txt and sitemap indexes (gzip supported) and returns up to 2000 URLs with lastmod dates.

Extract HTML tables from a page as JSON rows and records. All HTML tables on a page as JSON: header, rows, and records keyed by header when headers are unique.

Detect the technologies a website uses (CMS, frameworks, analytics, CDN). Detects CMS, e-commerce platform, JS framework, analytics, payments, CDN and server software from a page's HTML and headers.

Extract text from a PDF URL, page by page. Text of a PDF (up to 25 MB, first 50 pages) page by page, with document metadata and a hint when the PDF is a scan.

Domain health: DNS, SPF/DMARC, TLS certificate expiry, security headers. One-call domain audit: A/AAAA/MX/NS/CAA records, SPF and DMARC, TLS issuer and days to expiry, redirect chain and security-header score.

Convert Word, Excel, PowerPoint, PDF, EPUB or HTML to Markdown. Give a document URL (docx, xlsx, xls, pptx, pdf, epub, html, csv, json, xml; up to 30 MB) and get clean Markdown with headings, lists and tables kept.

Read text from an image (OCR), 7 languages. Optical character recognition on an image URL (png, jpg, webp, tiff, bmp, gif; up to 15 MB) with Tesseract. Languages eng, deu, fra, spa, ita, por, nld, combinable with + (e.g. eng+fra).

Measure audio loudness (EBU R 128 LUFS, true peak, loudness range). Two-pass EBU R 128 measurement of any audio or video file URL (up to 100 MB, 30 min): integrated LUFS, true peak, loudness range, and the gain needed for 11 named delivery targets.

Inspect a video: duration, size, rotation and what a storyboard would produce. Reads a video URL (up to 200 MB, 2 hours) and returns codec, size, display rotation, aspect ratio, duration and the storyboard a preset would produce.

DMARC report summary: reporter, date range, policy, record count. Lists what a DMARC aggregate report contains (reporter, report id, dates, published policy, records, messages) without evaluating it.

Inspect a map file: format, feature count, geometry types, extent, CRS. Reads a spatial file URL and returns its format, feature count, geometry types, extent, declared coordinate system and fields.

AI summary of any web page with key points and topics. Fetches a public page and returns an AI summary (short, medium or long), 3-7 key points and topics. Uses only facts on the page.

Contract check: proxy type, implementation, owner, verified source, capability flags. Inspects any EVM contract: detects EIP-1967 and EIP-1167 proxies and their implementation, reads the owner, checks Sourcify verification, and lists capabilities (mint, pause, blacklist, burn, upgrade) from the logic contract's bytecode.

AI extraction of named fields from any web page as JSON. Give a URL and comma-separated field names; get a JSON object with the best value per field, plus a list of records when the page lists many (products, people, events).

Normalise audio loudness to a broadcast or streaming target, file returned. Linear two-pass EBU R 128 loudness normalisation to a named target (ebu_r128, atsc_a85, ebu_r128_short, podcast, spotify, youtube, apple_music, amazon_music, deezer, aes_streaming, audiobook_acx) or custom lufs/true_peak. Returns the corrected file (wav, wav16, flac, mp3, m4a, opus) with before/after figures in X-Loudness-

Video scrub-bar storyboard: sprite sheets plus WebVTT thumbnail track (zip). Hover-preview thumbnails for a video player: sprite sheets capped at the device texture limit plus a WebVTT track whose cues use each frame's real timestamp, so they never drift. Handles phone rotation. Presets web_player, web_player_hd, mobile (WebP, 2048 cap), editor_filmstrip, contact_sheet.

DMARC aggregate report parser with real alignment checks (JSON or CSV). Reads a DMARC aggregate report URL (XML, .gz, .zip or the whole email). Computes SPF and DKIM alignment itself using the Public Suffix List, reports the receiver's verdict separately, flags disagreements, explains each failure in plain words, and rolls up pass rate, dispositions, failure modes and worst sources.

Convert and reproject map files (GeoJSON, shapefile, CSV, GPX, KML). Spatial file in, spatial file out: reads GeoJSON, zipped shapefile, CSV, GPX or KML and writes GeoJSON, CSV, WKT or zipped shapefile in any EPSG/PROJ/WKT system. Always longitude/easting first, reports the transformation and its accuracy in metres, and refuses data outside the target's area of use (spotting swapped lat/lon).

Text to natural speech (MP3 or WAV), 30 voices. Turns up to 1500 characters into natural speech with Google's Gemini TTS. 30 voices (e.g. Kore, Puck, Charon, Aoede). MP3 or WAV.

Web search with a grounded answer and source links (Google Search). Searches the live web with Google Search and returns a concise factual answer plus the source pages (title and real URL) and the queries used. Good for current facts, prices, news and docs.

Generate an image from a text prompt (PNG). Text-to-image with Google's Gemini image model. Aspect ratios 1:1, 3:2, 2:3, 3:4, 4:3, 4:5, 5:4, 9:16, 16:9, 21:9. Prompts refused by the model's safety filter are not charged.

YouTube titles, description, tags and hashtags for a video topic or transcript. Eight title options (with hook type and why each works), a description whose first lines work as the preview, 12 to 20 tags kept under YouTube's length limit, hashtags and chapter lines. Give a topic and optionally the transcript, audience and language.

Translate an SRT or VTT subtitle file, keeping every timestamp. Give the URL of an .srt or .vtt file (up to 800 cues) and a target language; get the same file back translated with identical timing, ready to upload.

Solana token safety check: mint and freeze powers, Token-2022 traps, holders, 0-100 risk score. Check a Solana token (SPL or Token-2022) before buying it. Reads the mint from the chain: who can mint more, who can freeze wallets, permanent delegate, transfer hooks and fees, default-frozen accounts, pausable, whether name and image can change, and top holders when available. 0-100 risk score with pl

Explain a Solana transaction in plain English: who paid whom, swaps, programs, red flags. Give a Solana transaction signature and get a plain-English summary: success or failure, fee, which accounts sent and received SOL and tokens, whether it looks like a swap, which programs ran (Jupiter, Raydium, Orca, Meteora, Pump.fun and more) and red flags such as token approvals, authority changes or unkno

Funny AI roast of a token, website or project, with a grade and a shareable one-liner. Give a token address (Base or Ethereum), a website or a project name and get a witty, harmless roast, a funny grade and a one-line quote to share. Styles: friendly, savage, pirate, shakespeare, hype. Roasts things, never people. Entertainment only.

Translate text into any language, keeping tone, names and formatting. Translates up to 4000 characters into any language (name or code), auto-detects the source language and flags ambiguities. Treats the text purely as data.

Sentiment, emotions, topics and toxicity of text, optionally sorted into your own labels. Analyses up to 4000 characters: language, sentiment and score (-1 to 1), up to 3 emotions, topics, toxicity and a one-line summary. Pass labels=a,b,c to sort the text into your own categories with a score for each.

Speech to text from an audio or video file URL (up to 5 minutes), with timestamps, SRT or VTT. Transcribes speech in an audio or video file (mp3, wav, m4a, mp4, webm and more; up to 80 MB and 5 minutes) with OpenAI Whisper running on our own servers. Auto-detects the language (99 languages). Returns text with segment timestamps, or ready-made SRT or VTT subtitles.

Thumbnail critique: score 0-100, readability at feed size, text, contrast, clickbait risk, fixes. Looks at a thumbnail image URL (jpeg, png or webp, up to 4 MB) the way a viewer sees it at feed size and returns a 0-100 score, whether it reads when small, the exact text in it, contrast, focal point, face and emotion, clickbait risk, strengths and concrete fixes. Pass the video title to check title

Video script: hook, sections with narration and visuals, outro, sized to your length. Writes a spoken script for a video of 1 to 15 minutes: a strong first-15-seconds hook, three to six sections each with narration and a visuals note, and an outro with a call to action. Choose style, audience and language.

Check a translated subtitle file against the original: meaning, grammar, timing, reading speed. Give the URLs of the original .srt or .vtt and its translation. A reviewer reads every cue pair and returns a 0-100 score, a verdict and only the cues that need attention (mistranslation, omission, grammar, spelling, unnatural phrasing) with a suggested correction, plus technical checks: cue count, timi