# Agent Tools by OpenKrill > Agent Tools by OpenKrill provides free Model Context Protocol (MCP) servers and streamable HTTP data APIs for AI coding assistants, autonomous agents, and developers worldwide: open jobs from company careers pages, app store reviews, X posts, and the per-tool servers listed below. Results are cached, so repeat requests are answered from the cache. Base URL: https://api.openkrill.app (this page: https://tools.openkrill.app) ## About OpenKrill OpenKrill builds Agent Tools: MCP servers and HTTP APIs that AI agents call for live public data. The tools are free to use, with no sign-up and no API key, within rate and daily limits. Packages with their own page, price and order button: Mailfix at https://tools.openkrill.app/mailfix, Partner program at https://tools.openkrill.app/partners, Dependency report at https://tools.openkrill.app/devdeps, Reference check at https://tools.openkrill.app/papers, Store Watch at https://tools.openkrill.app/storewatch, TikTok starter pack at https://tools.openkrill.app/tiktok-pack, HTS codes and landed cost at https://tools.openkrill.app/tradecodes, Cloudflare bill sheet at https://tools.openkrill.app/cfbill, Marketing brain starter at https://tools.openkrill.app/mktbrain, Vietnam quarterly tax kit at https://tools.openkrill.app/vi/khai-thue-quy (in Vietnamese), Shop quote MCP at https://tools.openkrill.app/shopquote, AI Build Day at https://tools.openkrill.app/build-day, Agent team setup at https://tools.openkrill.app/agent-team, Oddwatch at https://tools.openkrill.app/oddwatch. A free OpenClaw setup kit that splits one AI into 4 roles (lead, researcher, writer, checker), one Markdown file with no sign-up: https://tools.openkrill.app/ai-staff-kit/openclaw-4-roles-kit.md, https://tools.openkrill.app/vi/4-nhan-vien-ai/bo-4-nhan-vien-ai-openclaw.md (in Vietnamese). To reach us, email support@openkrill.app. Who we are: https://tools.openkrill.app/about. Contact and security reports: https://tools.openkrill.app/contact. What we keep: https://tools.openkrill.app/privacy. Terms of use: https://tools.openkrill.app/terms. ## Live tools One line per live MCP server: a link to its page on this site, which says what it does, who it is for and the price. Then a one-line summary and the MCP URL (streamable HTTP, no key needed). Every tool is free within the limits. The full list is at https://tools.openkrill.app/tools. Every server's agent skill, with its SKILL.md, MCP URL and page, is in one JSON file at https://tools.openkrill.app/skills.json. - [Agent Tools API (`get_transcript`, `list_jobs`, `get_app_reviews`, `read_x_post`)](https://tools.openkrill.app/#tools): Open jobs from company careers pages, app store reviews and public X posts. Free. MCP: https://api.openkrill.app/mcp - [Company Jobs Finder](https://tools.openkrill.app/tools/company-jobs): See who is hiring at the companies you name, and how a job fits you. Free. MCP: https://jobs.openkrill.app/mcp - [Startup Name Check](https://tools.openkrill.app/tools/names): Check if a .com is still free for your business name. Free. MCP: https://names.openkrill.app/mcp - [Recall Check](https://tools.openkrill.app/tools/recalls): Find US recalls for your car, a product or a food. Free. MCP: https://recalls.openkrill.app/mcp - [Citation Checker](https://tools.openkrill.app/tools/citations): Check that references are real and that a paper was not retracted. Free. MCP: https://citations.openkrill.app/mcp - [Store Price Tracker](https://tools.openkrill.app/tools/store-prices): Watch a Shopify store's prices and see what changed. Free. MCP: https://stores.openkrill.app/mcp - [Business Days & Workday Calc](https://tools.openkrill.app/tools/business-days): Count working days and public holidays for a country or region. Free. MCP: https://businessdays.openkrill.app/mcp - [Tariff & HS Code Duty Finder](https://tools.openkrill.app/tools/tariffs): Find candidate tariff codes and the duty for UK and US imports. Free. MCP: https://tariffs.openkrill.app/mcp - [EU Business & VAT Checker](https://tools.openkrill.app/tools/eu-business): Check an EU VAT number and look up a company's legal record. Free. MCP: https://eubusiness.openkrill.app/mcp - [Regulation Reader](https://tools.openkrill.app/tools/rules): Read US federal regulations and new proposals in the Federal Register. Free. MCP: https://rules.openkrill.app/mcp - [Site SEO & AI Crawler Audit](https://tools.openkrill.app/tools/site-check): See which AI crawlers a site lets in, missing share tags and redirects. Free. MCP: https://sitecheck.openkrill.app/mcp - [Travel Check](https://tools.openkrill.app/tools/travel): Read official travel advice and live US airport delays. Free. MCP: https://travel.openkrill.app/mcp - [Food Facts](https://tools.openkrill.app/tools/food): Read what is in a packaged food from its barcode, and check allergens. Free. MCP: https://food.openkrill.app/mcp - [Car Check](https://tools.openkrill.app/tools/cars): Decode a VIN and see recalls and fuel use for a US car. Free. MCP: https://cars.openkrill.app/mcp - [Document Tools](https://tools.openkrill.app/tools/docs): Count words and SMS segments, and make calendar and CSV files. Free. MCP: https://docs.openkrill.app/mcp - [Tax Numbers](https://tools.openkrill.app/tools/tax): Get US federal tax figures, each with the IRS document it came from. Free. MCP: https://tax.openkrill.app/mcp - [Catalog QA Desk](https://tools.openkrill.app/tools/catalog-qa): Check a product CSV before you import it, and find Merchant Center problems. Free. MCP: https://catalogqa.openkrill.app/mcp - [Dev Facts Check](https://tools.openkrill.app/tools/dev-facts): Get end-of-life dates, license facts, browser support and RFC status. Free. MCP: https://devfacts.openkrill.app/mcp - [CVE Risk Check](https://tools.openkrill.app/tools/cve): See how bad a CVE is and which dependencies to fix first. Free. MCP: https://cve.openkrill.app/mcp - [Domain Mail Check](https://tools.openkrill.app/tools/mailcheck): Find out if a domain's email will reach the inbox, and how to fix it. Free. Free web checker: https://tools.openkrill.app/spf-dkim-dmarc-check. MCP: https://mailcheck.openkrill.app/mcp - [SEO Audit](https://tools.openkrill.app/tools/seo-audit): Get a ranked list of a site's SEO problems, each with Google's rule. Free. MCP: https://seoaudit.openkrill.app/mcp - [Transcript Kit](https://tools.openkrill.app/tools/transcript-kit): Transcribe your own audio or video and get captions. Free. MCP: https://transcript.openkrill.app/mcp - [Trade Chart](https://tools.openkrill.app/tools/trade-chart): Draw a candlestick chart of a ticker with the indicators you name. Free. MCP: https://chart.openkrill.app/mcp - [Send Out](https://tools.openkrill.app/tools/send-out): Turn an answer into an email draft, calendar event or share link you send yourself. Free. MCP: https://sendout.openkrill.app/mcp - [Video Grab](https://tools.openkrill.app/tools/videograb): Find freely licensed videos, with the license from the source record. Free. MCP: https://videograb.openkrill.app/mcp - [Remote Jobs by Country](https://tools.openkrill.app/tools/remote-jobs): Find remote jobs you can be hired for from your country, each with the line of the post that says so. Free. [Browse open remote jobs](https://tools.openkrill.app/remote-jobs). MCP: https://remotejobs.openkrill.app/mcp - [Page Fetch](https://tools.openkrill.app/tools/page-fetch): Get a public page as clean Markdown with its links, after a robots.txt check. Free. MCP: https://pagefetch.openkrill.app/mcp - [Post Stats](https://tools.openkrill.app/tools/post-stats): Read the public views, likes and comments of YouTube and TikTok posts, and rank them. Free. MCP: https://poststats.openkrill.app/mcp - [CertPacket](https://tools.openkrill.app/tools/certpacket): Find the CPSC rules and certificate a product needs, accepted labs in Vietnam and China, and gaps in a lab report. Free. MCP: https://certpacket.openkrill.app/mcp - [Photo Sheet](https://tools.openkrill.app/tools/photosheet): Turn a receipt, notebook page, chat screenshot or voice note into a CSV table, with the sums checked. Free. MCP: https://photosheet.openkrill.app/mcp - [Belief Poll](https://tools.openkrill.app/tools/belief): Turn one question into belief percentages from X posts and Reddit comments, each percent linked to its posts. Free. MCP: https://belief.openkrill.app/mcp - [IndexNow Submit](https://tools.openkrill.app/tools/indexnow): Check your IndexNow key file, then send your sitemap to Bing, Yandex and 4 more engines. Free. MCP: https://indexnow.openkrill.app/mcp - [BillPeak](https://tools.openkrill.app/tools/billpeak): Recheck a small business power bill against the utility's published rate and find a cheaper plan. Free. MCP: https://billpeak.openkrill.app/mcp - [Settlement Claim Finder](https://tools.openkrill.app/tools/settlements): Find open US class action settlements that fit where you live and what you bought. Free. MCP: https://settlements.openkrill.app/mcp - [Share Link Resolver](https://tools.openkrill.app/tools/sharelink): Turn a public Google Drive link into a direct download URL for curl or Aria2, and list a Baidu share. Free. MCP: https://sharelink.openkrill.app/mcp - [E-Commerce Store Enricher](https://tools.openkrill.app/tools/store-enricher): See a store's platform, product count, prices, apps and contacts in one answer. Free. MCP: https://storeenricher.openkrill.app/mcp - [Associate Product Finder](https://tools.openkrill.app/tools/associate-links): Search Amazon on your own Creators API key, get links with your tag, and rank picks for an Amazon Live carousel. Free to use, with full results, on your own key. MCP: https://associatelinks.openkrill.app/mcp - [Sound Vocabulary](https://tools.openkrill.app/tools/sound-vocab): Describe a click, ding or whoosh and get a WAV and Web Audio code, named by UCS. CC0. Free. MCP: https://soundvocab.openkrill.app/mcp - [Tu Vi Chart](https://tools.openkrill.app/tools/tuvi): Cast a Tu Vi (Zi Wei Dou Shu) birth chart from a birth date and time, with readings that cite the classic. Free. Chart page: https://tools.openkrill.app/vi/la-so-tu-vi (in Vietnamese); Tu Vi by birth date: https://tools.openkrill.app/vi/tu-vi-theo-ngay-sinh (in Vietnamese). MCP: https://tuvi.openkrill.app/mcp - [Hiring Signals](https://tools.openkrill.app/tools/hiring-signals): Brief a company's hiring and watch companies for new and closed roles. Free. [Fastest-hiring companies this week](https://hiring.openkrill.app/leaderboard). MCP: https://hiring.openkrill.app/mcp ## Set up a tool in one step The name of a tool is the last part of its page address, such as site-check for https://tools.openkrill.app/tools/site-check. For a server that needs no key, that name and the tool's MCP URL make one command that connects it. A server that needs a key has its own steps on its page. Start a new agent session after it. - Claude Code: `claude mcp add --transport http site-check https://sitecheck.openkrill.app/mcp` - Hermes Agent: `printf 'n\ny\n' | hermes mcp add site-check --url https://sitecheck.openkrill.app/mcp`. The two answers are for its setup questions: no key, and enable all tools. - Any other agent: read the tool's page as Markdown (its page address plus .md), which has the steps for each client. A user can ask an agent: "set up Site SEO & AI Crawler Audit from https://tools.openkrill.app/llms.txt". ## Crawlers robots.txt allows every crawler on every page and names these in their own allow group: Googlebot (Google Search and Brave Search), Bingbot (Bing, and the Bing index behind Yahoo and DuckDuckGo results), DuckDuckBot (DuckDuckGo), Slurp (Yahoo), Applebot (Apple Spotlight and Siri), YandexBot (Yandex), OAI-SearchBot (ChatGPT search), PerplexityBot (Perplexity), ClaudeBot (Claude model training), Claude-SearchBot (Claude search), ChatGPT-User (ChatGPT, for a person's request), Claude-User (Claude, for a person's request), Perplexity-User (Perplexity, for a person's request), Google-Extended (Gemini), Amazonbot (Alexa and Amazon answers). Brave Search follows the Googlebot rules. ## Products - [Agent Tools](https://tools.openkrill.app/): free MCP servers and data APIs for AI agents and developers. - [KrillData](https://data.openkrill.app): an MCP server and REST API that gives AI agents web and social data. - [AI instructions](https://tools.openkrill.app/ai-instructions): how an assistant should read this site. - [Prices](https://tools.openkrill.app/pricing): every tool is free (0 USD) within its limits, and Stillchat is free. - [Agent Tools vs Smithery, Composio and Apify](https://tools.openkrill.app/vs): free plans, cheapest paid plans, sign-up before the first call and who hosts the MCP server, with sources read 2026-10-19. - [Stillchat](https://tools.openkrill.app/stillchat): a free iPhone and iPad app (iOS 26) that controls the AI coding agents on your own Mac. Pair by QR code, then approve, reject and steer Claude Code, Codex, ChatGPT and OpenRouter agents over an end-to-end encrypted relay. Not on the App Store yet. Support: https://tools.openkrill.app/stillchat/support. Privacy: https://tools.openkrill.app/stillchat/privacy. - [llms-full.txt](https://tools.openkrill.app/llms-full.txt): the longer text of the pages this file links to. ## Tools Each tool has an HTTP name (used in the route) and an MCP name (used with tools/call). - `yt-transcripts` (MCP: `get_transcript`): coming soon, not listed by `/v1/tools` or `tools/list` yet. The full transcript of one YouTube video, given a URL or an 11-character video ID. Output is timestamped segments (default), plain text, SRT or VTT. Optional `languages` in priority order and `translateTo`. A video without captions comes back as a result with status "error" and a code, not as an HTTP error. Cached for 30 days. - `job-boards` (MCP: `list_jobs`): open jobs at companies on Greenhouse, Lever, Ashby, Workable or SmartRecruiters. Input is `companies`, careers page URLs or "ats:slug" tokens such as "greenhouse:airbnb". Optional filters: `keyword`, `location`, `remote`, `department`, `postedSince`, `maxJobsPerCompany`. A board that cannot be read comes back as a record of type "error". Cached for 1 hour. - `app-reviews` (MCP: `get_app_reviews`): app metadata and recent reviews for Google Play and iOS App Store apps, given store URLs, Apple numeric ids or Play package names. Optional `countries`, `languages`, `maxReviewsPerApp`, `sinceDate` and `sort`. Apple's public review feed is sometimes empty; the result then holds the app row and an error row with code "unavailable". Cached for 6 hours. - `x-posts` (MCP: `read_x_post`): public X (Twitter) posts read through X's documented oEmbed endpoint (publish.x.com/oembed), given x.com or twitter.com URLs or post ids (up to 20), with no login. Each post comes back as text, author, day, language, the links in the text and whether it has media. X cuts posts longer than 280 characters in the embed. The embed has no counts, media files, quoted post, replies or thread, so the tool returns none of them. Since 2026-10-23 it uses no guest API, syndication host or mirror: X's terms ban scraping and x.com/robots.txt disallows every path. A post X will not embed (deleted, protected, suspended or age-restricted) comes back as a record of type "error" with code "unavailable". Only the posts you ask for are read, and nothing is kept past the cache time. Cached for 900 seconds (15 minutes). ## HTTP API - `GET /v1/tools`: list the tools with their input schemas. - `GET|POST /v1/tools/{name}/run`: run a tool. GET takes flat query parameters (repeat a key for arrays). POST takes a JSON body. - `GET /v1/jobs/{id}`: result of a slow request that was answered with HTTP 202. - `GET /v1/usage`: needs an API key. Billable items delivered to your API key this month, by charge event: reviews, jobs, and transcripts with status `ok`; error rows are free (`?period=YYYY-MM` for another month). No API key is needed: a call without one runs as an anonymous caller, limited to 20 requests a minute and 200 tool calls a day per network (the day resets at 00:00 UTC). Past a limit the API answers HTTP 429 with `Retry-After` in seconds and an error that says how to lift it. A spent keyless `list_jobs` call points to https://jobs.openkrill.app, which is not on that cap. An API key lifts those limits: send it as the `x-api-key` header, or as `Authorization: Bearer ` (60 requests a minute per key). A key that is wrong or disabled gets HTTP 401 and is never treated as anonymous. Keys come from support@openkrill.app. ## Paid calls Paid calls are off today: every tool runs on the free limits above. Paid calls are opt-in: a call that carries no payment stays free and unchanged, even with paid mode on, and `initialize` and `tools/list` never change. Until paid mode is on, `GET https://api.openkrill.app/v1/paid-tier` and `GET https://api.openkrill.app/v1/credits` answer 404 `payments_off`. - Paid tier: one x402 v2 payment in USDC on Base (`eip155:8453`) buys one call that skips the free per-minute and daily limits. `GET https://api.openkrill.app/v1/paid-tier` (or `/paid-tier` on a plugin) names the price and returns the `PaymentRequired` to sign. Over MCP put the signed payment in `params._meta["x402/payment"]` of the `tools/call`; the settlement comes back in `result._meta["x402/payment-response"]`. Over HTTP send it in the `PAYMENT-SIGNATURE` header; the settlement comes back in `PAYMENT-RESPONSE`. - A free-limit refusal keeps its status and code and, while paid mode is on, adds one sentence that points at the paid tier. A call that fails is not settled, so it costs nothing. - Credits: a key that holds prepaid credits draws the price of one call from them and skips the per-key limit; a key without credits runs free. `GET https://api.openkrill.app/v1/credits` shows the balance and the ways to pay. To buy a pack, POST the x402 pay URL it lists (it starts with `https://api.openkrill.app/v1/credits/x402` and names a credit pack). That POST answers 402 with a `PAYMENT-REQUIRED` header (and the same challenge in the body). Sign it with an x402 client and repeat the POST with the `PAYMENT-SIGNATURE` header. Send your `x-api-key` to top up that key; without one, a new key is created and returned once in the answer. - A payment is used once: repeating the same signed payment adds nothing. A failed run gives the call's credit back. - PayPal credit packs, when that rail is on: POST /v1/credits/paypal/order with x-api-key, open approveUrl, then POST /v1/credits/paypal/capture with the same key and the order id. A confirmed webhook credits the same capture once. The rail is absent until its secrets are set. - Polar credit packs, when that rail is on: POST /v1/credits/polar/order with x-api-key, open checkoutUrl and pay. Polar's signed order.paid webhook adds the pack to that key once; GET /v1/credits shows the balance. The rail is absent until its keys are set. Every run response carries `x-cache: HIT`, `MISS` or `STALE`, and a `Server-Timing` header. A request that takes longer than the server's wait budget is answered with HTTP 202, a job id and a poll URL while the work continues; the finished result is then served from the poll URL and from the cache. Rate limits are applied per network without a key and per API key with one, and answered with HTTP 429 and `Retry-After`. Example: ``` curl "https://api.openkrill.app/v1/tools/job-boards/run?companies=greenhouse:airbnb&keyword=engineer&maxJobsPerCompany=5" ``` With a key, for higher limits: ``` curl -H "x-api-key: $KEY" "https://api.openkrill.app/v1/tools/job-boards/run?companies=greenhouse:airbnb&keyword=engineer&maxJobsPerCompany=5" ``` ## MCP Endpoint: https://api.openkrill.app/mcp (streamable HTTP, stateless: POST one JSON-RPC message or a short batch, get one JSON reply). Tools listed by `tools/list` today: `list_jobs`, `get_app_reviews`, `read_x_post` (`get_transcript` follows once YouTube transcripts are live). MCP calls wait for the result instead of returning a job id. No key is needed; the same anonymous limits apply, and the same optional API key header lifts them. A limit is reported as a tool error whose text starts with `rate_limited` or `anonymous_daily_limit`. ## Connect with OAuth MCP clients that support OAuth (ChatGPT and Claude "Connect", or any client following the MCP authorization spec) can get an access token on their own; there is no key to create and no sign-in. - MCP endpoint: https://connect.openkrill.app/mcp. Without a token it answers HTTP 401 with a `WWW-Authenticate` header that points to `https://connect.openkrill.app/.well-known/oauth-protected-resource/mcp`. - Authorization server metadata: https://connect.openkrill.app/.well-known/oauth-authorization-server. It lists dynamic client registration (RFC 7591, `POST /oauth/register`), the authorization endpoint (authorization code with PKCE S256, approved automatically) and the token endpoint (refresh tokens rotate). - Scope: `tools` (run the tools above); add `offline_access` for a refresh token. - The access token also works as `Authorization: Bearer ` on https://api.openkrill.app/mcp and `/v1/tools/{name}/run`, with limits counted for your client. - Registration and authorization are limited per network address; calls with a token are limited per client. ## ChatGPT plugins Public MCP servers with no API key and no key option, one per plugin, for ChatGPT and other MCP clients. Calls are limited per network and per day; only aggregate counts are kept: calls per tool per day, plus counts of refused calls by reason, tool failures by kind, tool latency and the client family (for example chatgpt, claude, cursor, vscode or other) that connected, never an address, User-Agent, argument or answer, and a scheduled test call of each plugin is recorded as pass or fail (https://tools.openkrill.app/privacy). - Company Jobs Finder: https://jobs.openkrill.app/mcp. Four tools. `find_company_jobs`: open jobs at up to 5 named companies (names or careers page links) from Greenhouse, Lever, Ashby, Workable and SmartRecruiters, with `keyword`, `location`, `remote`, `department`, `postedSince` and `withSalary` filters. `rank_jobs_for_profile`: ranks the open jobs at up to 3 companies against a short profile the caller supplies (target and recent titles, years, level, skills, locations, remote), with a why-fit, gaps and a level-fit verdict per job from fixed rules, no AI model; the profile is used in memory only and is never stored, cached or logged. `job_fit_brief`: one posting's must-haves, nice-to-haves, level signals, stated process facts and a tailoring checklist drawn from general recruiter practice (no guarantee). `profile_keyword_check`: compares profile text the user pastes (never fetched from LinkedIn) with the title, skill, tool and seniority words recruiters commonly search for target titles or one posting, and returns what is present or missing, an example boolean search string and a short career-story tip, from fixed rules and no AI model; the text is used in memory only and is never stored, cached or logged. All read-only, public job data only. - Startup Name Check: https://names.openkrill.app/mcp. One tool, `check_domain_availability`: whether the domains for up to 5 names or domains are registered (Cloudflare Registrar domain-check when configured, otherwise RDAP; default endings com, net, org, ai, app, dev; `tlds` to choose others), with registration and expiry dates when published, a registration cost when the registrar returns one, and never owner details. Every answer names its source and the date it was read. Indicative, not a reservation; trademarks are not checked. Read-only. - Recall Check: https://recalls.openkrill.app/mcp. Five tools: `check_vehicle_recalls` (year, make, model), `decode_vin`, `search_product_recalls` (CPSC consumer products), `search_food_drug_recalls` (FDA food and drug) and `search_pet_food_recalls`. US data only, not a safety guarantee. Read-only. - Citation Checker: https://citations.openkrill.app/mcp. One tool, `check_references`: up to 10 references or DOIs at a time, each checked against Crossref, the DOI system (doi.org) and OpenAlex, with a verdict of found, mismatch or not_found and the paper's retraction status. Bibliographic details only. Read-only. - Store Price Tracker: https://stores.openkrill.app/mcp. Two tools, `check_store_prices` and `get_product_price_history`: reads the public product feed (/products.json) of Shopify stores, up to 5 stores at a time, and reports products, prices, variants and availability plus new and removed products and price and stock changes since the previous daily snapshot. A store that is asked about is added to a daily snapshot list so product price history builds up. Honors robots.txt, identifies itself as StorePriceTracker and backs off when a store rate limits it. Public product data only. - Business Days & Workday Calc: https://businessdays.openkrill.app/mcp. Also called OpenKrill Business Days, Workday Calculator and Bank Holiday Calendar. Three tools: `count_business_days` (working days, workdays, office days and PTO days between two dates for a country or region), `add_business_days` (the date a number of business days before or after a date, including an SLA due date) and `list_holidays` (bank holidays and school holidays from Nager.Date and OpenHolidays). Each answer states how days were counted: weekend days, which holidays applied, and whether the start and end dates were included. Public holiday calendars only, not a legal deadline calculator. Read-only. Country pages with the tool's own 2026 counts: https://tools.openkrill.app/business-days/germany-2026, https://tools.openkrill.app/business-days/england-2026, https://tools.openkrill.app/business-days/united-states-2026, https://tools.openkrill.app/business-days/france-2026, https://tools.openkrill.app/business-days/vietnam-2026, https://tools.openkrill.app/business-days/ontario-2026. - Tariff & HS Code Duty Finder: https://tariffs.openkrill.app/mcp. Also called HS Code Finder and Customs Code Lookup. Two tools: `find_tariff_codes` (candidate tariff codes for a described product, from the UK Trade Tariff or the US Harmonized Tariff Schedule) and `get_duty_rate` (the import duty on a code; for the UK also VAT, the preferential rate for an origin country and other measures). Every code is a candidate, not a binding classification; chapter 99 additional duties for the US are not included. UK and US only. Read-only. - EU Business & VAT Checker: https://eubusiness.openkrill.app/mcp. Also called EU VAT Checker and VIES VAT Validator. Three tools: `check_vat_number` (whether an EU or Northern Ireland VAT number is valid, with the registered name of a legal entity where the member state publishes it and a VIES consultation number on request), `find_company_lei` (the Legal Entity Identifier of a company, from GLEIF) and `lookup_company_registry` (the French Sirene and Norwegian Enhetsregisteret company registers). Names are shown only for legal entities; sole proprietors are left out. Read-only. - US Federal Regulations (CFR): https://rules.openkrill.app/mcp. Also called CFR Lookup and eCFR Search. Three tools: `get_cfr_section` (the current text, authority and amendment history of a section of the US Code of Federal Regulations, from the eCFR), `search_cfr` (which sections of the current regulations cover a topic) and `search_federal_register` (rules, proposed rules, notices and presidential documents, with comment deadlines and effective dates). General information, not legal advice; the eCFR is an unofficial edition; US federal regulations only. Read-only. - Site SEO & AI Crawler Audit: https://sitecheck.openkrill.app/mcp. Also called Site Check and Redirect Chain Checker. Eight tools: `check_ai_crawler_access` (whether a site's robots.txt, headers and redirects let 14 named AI crawlers fetch a page), `check_page_tags` (title, description, canonical, language, Open Graph and Twitter tags, and which are missing) and `trace_redirects` (each redirect hop, the final address, https upgrade and security headers), plus five more: `landing_page_check` (conversion checks on one landing page from its HTML: headline, value proposition, call to action, link targets, forms, trust signals, next step, mobile viewport, page weight, contact path; each finding has evidence and a fix; wording follows the MIT-licensed coreyhaines31/marketingskills conversion guidance), and three for online stores on any platform (Shopify fields named where they apply): `product_page_seo` (one product page's title, description, h1, canonical link and Product structured data), `ai_readiness_check` (whether a domain publishes llms.txt, agents.md, the Shopify UCP file, a sitemap and AI crawler rules) and `link_sample_check` (broken pages and redirects in up to 40 sitemap addresses, using HEAD requests), and `exposure_check` (for a domain or host the caller proves they own by a DNS TXT record or a well-known file: answer first, exposed or clean, the top 3 risks with a fix each, then the evidence; dangling subdomains from certificate transparency and DNS, leaked files and debug pages from a fixed list of 8 paths, default install pages and the https redirect; no port scans, no person, user name or email lookups). The store tools and the landing page check read robots.txt first and skip pages it disallows for Agent Tools. Only public http and https sites are fetched; private, local and internal addresses are refused, also after a redirect. Apart from the owner-proven exposure check, it reports what a site declares, not a security scan or access control. Read-only. - SEO Audit: https://seoaudit.openkrill.app/mcp. Five tools: `audit_site` (a bounded crawl of up to 25 pages of one named site from its homepage and sitemap, robots.txt respected, returning failed SEO rules grouped by priority with the affected pages, the fix and the Google Search Central source link, including comparisons across the pages read: near-identical templated pages, pages with little text of their own, hreflang links not returned, sitemap pages nothing links to), `audit_page` (the page rules for one address), `local_ai_readiness` (pass, warn or fail for a local business site: a page per service, location or service-area pages, visible prices, FAQ, LocalBusiness schema, NAP consistency, on-site proof, a one-line description and recent dated content, each with the evidence URL and a fix, plus a Yext study note on what Gemini, OpenAI and Anthropic lean on), `check_ai_instructions` (whether the site publishes llms.txt, llms-full.txt and an AI or about-for-AI page, whether the home page links to them, and whether the file matches the site's own name, products and facts) and `generate_ai_instructions` (a draft llms.txt from those public pages for the owner to review; nothing is published). The AI instructions tools cite https://x.com/chris_nectiv/status/2105655128162468334 (2026-10-01) as the reason they exist. They do not query AI engines and do not promise an effect on AI Overviews. local_ai_readiness does not query those engines or fetch directories. Limited per day and per network. An optional per-page judgement of purpose, specificity and competing pages is off by default, limited per day and per network, and falls back to rule results with a note. Only the named public site is crawled; private, local and internal addresses are refused. Does not measure rankings, traffic or backlinks. Read-only. - Travel Check: https://travel.openkrill.app/mcp. Three tools: `get_travel_advice` (a country by name or ISO code: UK FCDO alert and US State Department level with dates and links, optionally the UK entry-requirements, safety or health text), `get_airport_status` (FAA delays for US airports plus METAR and TAF weather, by IATA or ICAO code) and `compare_advisories` (up to 5 countries side by side). UK advice is for British citizens and US advice for US citizens. Not medical or safety advice, not for flight planning. Read-only. - Food Facts: https://food.openkrill.app/mcp. Four tools: `lookup_barcode` (EAN or UPC barcode: ingredients, listed allergens, Nutri-Score, NOVA group, nutrition per 100 g, data-quality note), `search_products` (by name or brand, optionally a country), `compare_products` (up to 5 barcodes ranked by sugars, salt or kcal per 100 g) and `check_allergens` (the entry's listed allergens against ones to avoid). Data from Open Food Facts (ODbL), community-contributed: check the physical label. Not dietary or medical advice. Read-only. - Car Check: https://cars.openkrill.app/mcp. Four tools: `decode_vin` (17-character VIN: year, make, model, trim, engine, with NHTSA's error codes), `get_vehicle_report` (VIN or year, make, model: NHTSA crash ratings, recall count, complaint count and top components, EPA fuel economy), `list_recalls` (year, make, model) and `get_fuel_economy` (EPA mpg or mpge, range, annual fuel cost). Data from NHTSA and the US EPA. US vehicles only; recalls and complaints are by model, not VIN; complaints are not confirmed defects. Not a valuation, history report or safety advice. Read-only. - Document Tools: https://docs.openkrill.app/mcp. Four tools, `count_text`, `make_calendar_file`, `make_csv_file` and `translate_page`: counts characters, words, sentences, SMS segments and reading time (text is never stored), writes RFC 5545 calendar files and RFC 4180 CSV files as text plus a download link that works for 24 hours and is then deleted, and translates a public webpage or pasted text into a target language while keeping code blocks and URLs intact. - Tax Numbers: https://tax.openkrill.app/mcp. Five tools, `get_tax_parameters`, `get_mileage_rate`, `tax_deadlines`, `estimate_federal_tax` and `tax_duty`, give US federal tax figures for tax years 2025 and 2026 (brackets, standard deduction, long-term capital gains thresholds, Social Security wage base and self-employment tax, 401(k) and IRA limits), the IRS standard mileage rates since 2022, filing and estimated-payment deadlines, a simple federal estimate with its assumptions and exclusions, and VAT, GST or US sales-tax registration rules and thresholds for a buyer country or US state. Every row names the IRS or tax-authority document it comes from and every answer says when it was last verified. Estimates are not tax advice; the amounts you give are never stored. - Domain Mail Check: https://mailcheck.openkrill.app/mcp. Five tools: `launch_score` (a domain you will send cold email from and optional DKIM selectors: a score from 0 to 100 and a verdict of ready, almost, not_ready or incomplete, from SPF, DKIM, DMARC, MX, domain age from the public RDAP record and name pattern; each rule is kind standard, from an RFC or a published sender requirement, or kind opinion, common advice that never makes the verdict not_ready, and each links its source; it returns the first fixes in order and lists what it cannot test; it sends no message and cannot see sender reputation or warm-up), `delivery_check` (a domain, an optional sending IPv4 address and optional DKIM selectors: one verdict on whether Gmail and Outlook will likely accept the mail, the top 3 causes with one fix each, SPF for that IP, reverse DNS, DKIM key length, DMARC and 6 free blocklists, each rule with its official source link; it sends no test message and cannot see sender reputation), `check_email_auth` (a domain and optional DKIM selectors: SPF policy and DNS lookup count, DMARC policy, percentage and report address, DKIM per selector, MX hosts and mail provider, a grade from A to F, issues and fixes), `check_dns_records` (A, AAAA, MX, TXT, CAA and NS records with TTLs and the mail provider named by the MX hosts) and `lookup_domain_registration` (registered or not, registrar, created, updated and expiry dates, status flags and name servers from the registry's RDAP service). DNS comes from Cloudflare DNS over HTTPS with Google Public DNS as a fallback; registration data from IANA's RDAP directory and the TLD registry. Registrant and contact details are never returned. It checks the exact domain given and does not list subdomains; to check whether a startup name is free use Startup Name Check. - CVE Risk Check: https://cve.openkrill.app/mcp. Four tools: `check_cve` (a CVE id: priority exploited, likely, routine or unknown, CVSS score and severity, CISA Known Exploited Vulnerabilities status with due date, EPSS probability, a patch or advisory link and the NVD and EPSS as-of dates), `check_cves` (up to 10 CVE ids, most urgent first, with a count per priority), `list_recent_kev` (CVEs CISA added to its exploited list in the last 1 to 90 days, with an optional product word) and `triage_dependencies` (a lockfile, package.json, requirements.txt or package list: the 3 vulnerable packages to upgrade first, each with its priority, CVE ids and fixed version). Sources are the National Vulnerability Database, OSV.dev, the CISA exploited list and FIRST EPSS, no key; NVD records are cached for hours and a `rate_limited` status says when to retry. It cannot tell whether your code reaches a vulnerable function. - Catalog QA Desk: https://catalogqa.openkrill.app/mcp. Four tools: `csv_import_preflight` (a Shopify product CSV checked before import, with blocking problems and warnings by row), `merchant_feed_check` (Google Merchant Center rules for barcodes, titles and descriptions), `listing_text_audit` (missing image alt text, copied and thin descriptions, title problems) and `listing_fix_sheet` (an advisory CSV of corrected alt text and titles). The last three take a CSV or a store domain; a domain is read once, from the public /products.json after robots.txt, identifying itself as CatalogQADesk, up to 1250 products. Large stores may block that feed, so a CSV is the dependable input. Nothing is stored, no personal data is read, and nothing is changed in the store. Read-only. - Dev Facts Check: https://devfacts.openkrill.app/mcp. Six tools, `get_eol_status`, `check_runtime_eol`, `check_license`, `check_browser_support`, `lookup_rfc` and `get_release_notes`: end-of-life and supported release cycles of a software product (endoflife.date), runtime version pins in a repo or manifest checked for end-of-life status, SPDX license identifier with OSI and FSF status, Baseline status and first browser versions for a web feature (webstatus.dev), an RFC by number or title words with the RFCs that obsolete or update it (RFC Editor and IETF Datatracker), and release notes or changelog highlights for a software version. Every answer names its source and the date of the data. Not for package vulnerabilities; license answers are not legal advice. - Trade Chart: https://chart.openkrill.app/mcp. One tool, `chart`: a candlestick price chart for a public ticker (stocks, funds, a currency pair or a crypto pair), with volume and up to four indicators (sma, ema, rsi, macd, bollinger, atr, stochastic, vwap). Timeframes are 5m, 15m, 1h, 1d and 1w. The answer is a self-contained interactive HTML chart rendered by Vela (Apache-2.0, LuxAlgo; the attribution mark stays on; PineTS is not used) plus a plain-language read of the last price and the indicator values, with the source and as-of time. Prices come from Yahoo Finance public chart data, Coinbase public candles when a crypto pair is missing there, or ECB daily reference rates through Frankfurter for a currency pair with no candles. No key, no orders, no account. Not financial advice. Read-only. - Video Grab: https://videograb.openkrill.app/mcp. Three tools: `search_free_videos` (freely licensed video from Wikimedia Commons and the Internet Archive, with a download link, licence and attribution; Pexels and Pixabay stay off until their keys exist, and the answer says so), `own_video_export` (official steps to export videos you uploaded to TikTok, YouTube Studio, Instagram or your X archive) and `video_metadata` (a Commons file, an Internet Archive item, or a YouTube watch URL: public metadata only, the YouTube Data API when that key exists and oEmbed when it does not, never a download and never a caption file). A platform URL or a watermark request returns an explanation and no file. Read-only. - Transcript Kit: https://transcript.openkrill.app/mcp. Four tools: `transcribe_media` (an audio or video file you upload, or a direct https link to one, for a recording you own or may use: a timestamped transcript with SRT and VTT captions, transcribed by Cloudflare Workers AI Whisper), `search_transcripts` (passages and repeated themes across your own transcripts of the last 24 hours) and `suggest_clips` (the best 30 to 60 second stretches with start and end times and a hook line, as text) and `translate_subtitles` (a transcript of yours or pasted SRT/VTT text that you may use, translated into about 100 languages with the cue times kept, as SRT and VTT plus a dub script of spoken lines sized to each segment; up to 18,000 characters a call, not stored). Files are limited to 15 MB and 15 minutes, there are daily limits on minutes and characters, and links to video or social platform pages are refused. Transcripts are deleted after 24 hours and the media is never kept. - Send Out: https://sendout.openkrill.app/mcp. Seven tools that turn an answer into something you send yourself: `draft_email` (a mailto link and an .eml draft), `create_calendar_event` (a Google Calendar link and an .ics file), `create_whatsapp_link` (a wa.me click-to-chat link), `create_obsidian_note` (an obsidian:// note), `create_x_post` (an X share link with a 280-character check), `create_markdown_file` and `prepare_transfer` (a download card for any text export). Nothing is sent for you. - Remote Jobs by Country: https://remotejobs.openkrill.app/mcp. One tool, `find_remote_jobs`: remote jobs a person can be hired for from a given country, with an optional job family such as software, support or marketing. - Page Fetch: https://pagefetch.openkrill.app/mcp. One tool, `fetch_page`: one public web page as clean Markdown without menus, footers, scripts or cookie banners, plus its title, description, language, canonical address and up to 100 links; robots.txt is checked first. ## Install guides HTML pages generated from the plugin listings, so the tool names match `tools/list`. Each plugin has a page per client (chatgpt, claude, cursor, pi) with that client's install steps and one `tools/call` checked against the live MCP host. The check date is on the page. No key is required on the plugin hosts. - Index: https://tools.openkrill.app/integrations - Developer page (CVE Risk Check, Dev Facts Check, Citation Checker; dev-security install first; production limits): https://tools.openkrill.app/developers - Role guides (only roles the tools can actually serve): https://tools.openkrill.app/for - Glossary of MCP and agent terms: https://tools.openkrill.app/glossary - Sitemap: https://tools.openkrill.app/sitemap.xml - CVE Risk Check in Cursor: https://tools.openkrill.app/integrations/cve/cursor - Company Jobs Finder in ChatGPT: https://tools.openkrill.app/integrations/company-jobs/chatgpt - SEO Audit in Claude: https://tools.openkrill.app/integrations/seo-audit/claude Short questions answered, one page per server, each section opening with the direct answer and the date it was last checked (index: https://tools.openkrill.app/answers): - CVE Risk Check: https://tools.openkrill.app/answers/cve - Domain Mail Check: https://tools.openkrill.app/answers/mailcheck - Company Jobs Finder: https://tools.openkrill.app/answers/company-jobs A path is `/integrations//`, or `/answers/` for the short questions. Clients are `chatgpt`, `claude`, `cursor` and `pi`. These pages are not localized. hreflang is not set. ## Mirror Twin Before an agent acts on a task, it can ask for the task to be read back. A model on Cloudflare Workers AI returns the goal, the scope, what would be touched, the assumptions and the risk. Each part carries `basis`: `stated` when the person's own words back it, `filled_in` when they do not. `order` marks each clause of the task as `heard` or `missed`. `next_step.action` is `ask_person` while any part except the risk is `filled_in`: ask your person about each text in `next_step.ask_about`, send the answers as notes, and read again before you act. It is `go_ahead` once every part rests on their words. - Page: https://tools.openkrill.app/mirror - `POST https://tools.openkrill.app/mirror/read` with JSON `{"task": "...", "notes": [{"part": "...", "correction": "..."}]}`. The task is 8 to 1,000 characters; notes are optional, at most 6, and each names the part it corrects by that part's `text` from the last reading. Nothing typed is stored. Limits: 6 readings a minute and 30 a day per network. ## Inbox Document Sort Sort a business email or PDF you already have into purchase order, remittance advice, schedule change, vendor invoice, complaint or other. The answer has the fields, one sheet per class, a remittance check against an optional open-invoice CSV (matched, short_paid, missing) and a short briefing. Text PDFs are read here. A scan with no text layer needs ocr_text. Nothing is stored. - Page: https://tools.openkrill.app/docsort - MCP: `https://docsort.openkrill.app/mcp`, tool `sort_inbox`. - `POST https://tools.openkrill.app/docsort/sort` with form fields `text`, optional `open_invoices_csv` and `as_of`, or the same JSON the tool takes. The file is not stored. ## Asset library Original files for a web page or a video, all made by Agent Tools: 5 WebGL shaders, the Last Table design tokens, a star field texture, a poster scene, rail icons, 2 logos, a 120 BPM music bed and a nebula motion clip. The manifest records how each asset was made, its source file and its commit. When a task needs a background, an icon, a logo or music and must use no third-party files, pull from here instead of writing new ones. - Page: https://tools.openkrill.app/library - Manifest: https://tools.openkrill.app/library/manifest.json - Pull tool, one file for Node 18 or newer: `curl -O https://tools.openkrill.app/library/pull.mjs`, then `node pull.mjs search music for a film` and `node pull.mjs pull page-starter --out assets`. It checks every sha256 and writes `ASSETS.md` with the licence and origins. Output is JSON. - Licence: the shader code may be copied into any page; the other files are for Agent Tools pages and films until a public licence is chosen. ## Backend readiness checklist 16 things a backend needs before real users arrive, each with how to check it. The Site Check `exposure_check` tool reads 5 of them from outside on a host the caller proved they own: a bad request answered 500, rate-limit or 429 signals on the sign-in page, stack traces and tokens in error pages, a health endpoint that names its database, and a CORS policy that echoes any origin or uses a wildcard. Its answer has a `backend` object with `checks` (each labeled confirmed, needs_verification, not_checked or pass) and `ownerChecklist`, the items no outside request can see, each with how to verify it; those are never a pass or fail. It sends GET requests only and never tries to sign in or load the site. - Page: https://tools.openkrill.app/backend-checklist - MCP: https://sitecheck.openkrill.app/mcp, tool `exposure_check` ## Blog Posts built on our own tool results and measured numbers, each with its sources linked. RSS feed: https://tools.openkrill.app/blog/feed.xml. Index: https://tools.openkrill.app/blog - We checked the sites Google cites for "llms.txt generator": https://tools.openkrill.app/blog/llms-txt-generator-sites-checked (Markdown: https://tools.openkrill.app/blog/llms-txt-generator-sites-checked.md). Of the 6 vendor sites Google cited for "llms.txt generator" on 2026-10-08, 5 publish llms.txt, none publishes llms-full.txt, and no home page links to its file. - Can an agent read X posts without an API key? We measured our MCP tool: https://tools.openkrill.app/blog/read-x-posts-without-api-key (Markdown: https://tools.openkrill.app/blog/read-x-posts-without-api-key.md). Yes for single public posts: on 2026-10-23 our MCP tool read 4 of 4 posts through X's embed endpoint with text, author and day, but no counts, media files, quoted posts or replies. - Is my domain ready for cold email? Our own score before and after one DNS record: https://tools.openkrill.app/blog/cold-email-domain-check-before-after-dmarc (Markdown: https://tools.openkrill.app/blog/cold-email-domain-check-before-after-dmarc.md). A real before and after on our own domain: 80 (almost) with no DMARC record, 100 (ready) on the latest run after we added one TXT record, and the rules that still matter. - How old are the open jobs on Greenhouse careers pages? We measured 5 companies: https://tools.openkrill.app/blog/greenhouse-open-jobs-and-job-age (Markdown: https://tools.openkrill.app/blog/greenhouse-open-jobs-and-job-age.md). Measured posting ages from the public Greenhouse boards of 5 companies, and why an old posting is a question to ask, not proof of a ghost job. - Track a Shopify store's prices and stock from products.json: 3 real stores, measured daily: https://tools.openkrill.app/blog/shopify-products-json-price-tracker (Markdown: https://tools.openkrill.app/blog/shopify-products-json-price-tracker.md). Measured day-over-day changes on 3 real Shopify stores from their public products.json feed, plus a 1-request test for whether a site is Shopify. Agent Tools weekly sends these posts once a week: https://tools.openkrill.app/newsletter. To sign up, POST the form field `email` (application/x-www-form-urlencoded) to https://tools.openkrill.app/newsletter/subscribe; the answer is a 303 to /newsletter/check-your-inbox and the address joins only after its owner opens the confirm link in the mail. The confirmed subscriber count, as JSON with `subscribers`: https://api.openkrill.app/v1/newsletter/stats. Every sent issue, readable without subscribing: https://tools.openkrill.app/newsletter/archive (JSON: https://api.openkrill.app/v1/newsletter/archive). - We Checked 6 Product Names: 5 Failed on the .com, and npm Passed the One With a Live Product: https://tools.openkrill.app/blog/we-checked-6-names-for-a-new-tool (Markdown: https://tools.openkrill.app/blog/we-checked-6-names-for-a-new-tool.md). We ran no live trademark office search. The free USPTO search API answered 404 to our scripts, so we could not read it automatically. - Our Falco Sensor Went Silent for Hours While Still Running: 611 Connects in 85 Seconds From a Fresh Process: https://tools.openkrill.app/blog/our-falco-sensor-went-silent-for-hours-while-still-running (Markdown: https://tools.openkrill.app/blog/our-falco-sensor-went-silent-for-hours-while-still-running.md). Time to notice a silent sensor went from 3 hours, connects only, to about 1 hour, each sensor on its own. - A Node Job Hit Its 384 MB Cap on 11,933 Records: Keeping 4 Fields per Record Cut the Peak to 246 MB: https://tools.openkrill.app/blog/a-node-batch-job-hit-its-384-mb-memory-cap-on-11-933-records (Markdown: https://tools.openkrill.app/blog/a-node-batch-job-hit-its-384-mb-memory-cap-on-11-933-records.md). The judge step now keeps 4 fields per record: the key, the subject, the row and a due flag. - 9 Odd DNS Lookups Traced to Our Own Crawler: 8 CDN CNAME Targets and 1 Broken robots.txt Redirect: https://tools.openkrill.app/blog/9-odd-dns-lookups-traced-to-our-own-crawler (Markdown: https://tools.openkrill.app/blog/9-odd-dns-lookups-traced-to-our-own-crawler.md). A check that judges the name the program asked for, not the CNAME target, would not have fired on 8 of the 9. - Reading 8 App Store and Google Play Data URLs Against robots.txt by RFC 9309: 3 Allowed, 5 Disallowed: https://tools.openkrill.app/blog/reading-8-app-store-and-google-play-data-urls-against-robots (Markdown: https://tools.openkrill.app/blog/reading-8-app-store-and-google-play-data-urls-against-robots.md). Three rules did the work. Our tool sends no named product token, so the * group applies (section 2.2.1). - Reading YouTube Stats From a Link: One Test Video, Exact vs Rounded Numbers: https://tools.openkrill.app/blog/four-tools-that-scored-under-7-0-in-our-benchmark-and-what (Markdown: https://tools.openkrill.app/blog/four-tools-that-scored-under-7-0-in-our-benchmark-and-what.md). The verdict in our benchmark is an honest loss for our tool. It was one of 4 of our 55 tested tools that scored below 7.0/10 on real inputs. - Remote software jobs from Vietnam: what a country filter finds, and where it still guesses: https://tools.openkrill.app/blog/remote-jobs-by-country-company-jobs-finder-and-task (Markdown: https://tools.openkrill.app/blog/remote-jobs-by-country-company-jobs-finder-and-task.md). We ran our live endpoint and 3 named rivals on that exact input and timed wall-clock latency in ms. - An open model said it was Claude in 18 of 30 chats on our own GPU: https://tools.openkrill.app/blog/open-model-identity-test-claude-gpt (Markdown: https://tools.openkrill.app/blog/open-model-identity-test-claude-gpt.md). Ornith-1.5-9B never used its own name. Its 5 "other" answers all said Qwen, the family it is built on. - npm audit blocked 20 of 25 packages; an exploit-based gate blocked 3: https://tools.openkrill.app/blog/cutting-ci-dependency-blocks-from-20-to-3-packages-in-our (Markdown: https://tools.openkrill.app/blog/cutting-ci-dependency-blocks-from-20-to-3-packages-in-our.md). On a GitLab merge request, a commit pinning ejs 3.1.6 failed the pipeline in the secure stage. - 1 of 291 AI search answers named us: how to run the same share audit on your brand: https://tools.openkrill.app/blog/we-asked-5-ai-search-engines-90-tool-questions-and-1-of-291 (Markdown: https://tools.openkrill.app/blog/we-asked-5-ai-search-engines-90-tool-questions-and-1-of-291.md). For each tool we wrote 3 natural buyer or developer questions, so 90 prompts in all. - Is This CVE Actually Exploited? Checking CVE-2023-34362 Against OpenCVE and CIRCL: https://tools.openkrill.app/blog/cve-lookup-test-our-tool-vs-opencve-and-circl (Markdown: https://tools.openkrill.app/blog/cve-lookup-test-our-tool-vs-opencve-and-circl.md). Our tool tied OpenCVE on score and beat CIRCL on completeness. It was not the fastest: OpenCVE was 70 ms quicker. - Writesonic AI visibility trial: the first report said 16 percent (4 of 25 answers), the non-branded dashboard 2.7 percent (2 of 74): https://tools.openkrill.app/blog/writesonic-ai-visibility-trial-branded-prompts (Markdown: https://tools.openkrill.app/blog/writesonic-ai-visibility-trial-branded-prompts.md). Writesonic's first AI visibility report gave our site 16 percent, but some of the test prompts it used named our domain. The dashboard's non-branded view, a separate set of 74 answers, showed 2.7 percent. That's 2 of 74. - Synthesia free plan test: a 30-second avatar video in about 4 minutes, with garbled text in the first 5 seconds: https://tools.openkrill.app/blog/synthesia-free-plan-avatar-video-test (Markdown: https://tools.openkrill.app/blog/synthesia-free-plan-avatar-video-test.md). On Synthesia's free Basic plan the chat builder turned our 52-word script into a 30-second avatar video in about 4 minutes and let us download a watermarked 1080p file. The first 5 seconds showed garbled text that the builder's own check missed. - Murf free plan test: 17 of 20 tech terms read clearly, and the free plan cannot download the audio: https://tools.openkrill.app/blog/murf-free-plan-tech-terms-voice-test (Markdown: https://tools.openkrill.app/blog/murf-free-plan-tech-terms-voice-test.md). Murf's free plan read 17 of 20 developer terms clearly in our 149-word script. It likely said OAuth as oath and clipped dot env. It used 56 seconds of its 10 free minutes, counted by audio length, and the free plan does not let you download the audio. - Kapwing free plan test: auto subtitles got 140 of 141 words right in 12.7 seconds, then exported at 720p with a watermark: https://tools.openkrill.app/blog/kapwing-free-plan-auto-subtitles-test (Markdown: https://tools.openkrill.app/blog/kapwing-free-plan-auto-subtitles-test.md). Kapwing's free plan subtitled our 49-second clip in 12.7 seconds for 1 of 10 credits, with 140 of 141 words exact. The free export is capped at 720p and carries the Kapwing logo on every frame. - Gamma free plan test: a 1,412-word post became 10 slides in about 70 seconds, with all 58 numbers kept and 1 label changed in meaning: https://tools.openkrill.app/blog/gamma-free-plan-blog-post-to-10-slides (Markdown: https://tools.openkrill.app/blog/gamma-free-plan-blog-post-to-10-slides.md). Gamma's free plan turned our 1,412-word post into a 10-slide deck in about 70 seconds and kept all 58 numbers. But it turned load average into active cores and added words like rigorous that the post never used. - Kit trial test: an API broadcast went out 9 seconds after its send time and landed in the inbox: https://tools.openkrill.app/blog/kit-trial-api-broadcast-test (Markdown: https://tools.openkrill.app/blog/kit-trial-api-broadcast-test.md). On a new Kit account during its Creator trial, a broadcast scheduled through the Kit API went out 9 seconds after its set time and landed in the Gmail inbox. The free plan covers broadcasts to up to 10,000 subscribers, while sequences and automations are paid. - n8n Cloud trial test: a 10-item RSS run counts as 1 execution, and editor test runs count as 0: https://tools.openkrill.app/blog/n8n-cloud-trial-one-execution-per-run (Markdown: https://tools.openkrill.app/blog/n8n-cloud-trial-one-execution-per-run.md). On an n8n Cloud trial, a webhook run that read and reshaped 10 RSS items counted as 1 execution of 1,000, and 3 test runs from the editor counted as 0. The same job costs 1 execution on n8n and 11 credits on Make. - Make free plan test: one 10-item RSS run costs 11 credits, and an empty check still costs 1: https://tools.openkrill.app/blog/make-free-plan-rss-run-11-credits (Markdown: https://tools.openkrill.app/blog/make-free-plan-rss-run-11-credits.md). On the Make free plan, a scenario that read 10 RSS items and handled each one used 11 credits a run, and a watch trigger that found nothing still used 1 credit per check. So the 1,000 monthly credits cover an hourly empty check but not one every 15 minutes. - How to use Claude for dependency audits: 6 prompts on a real lockfile: https://tools.openkrill.app/blog/how-to-use-claude-for-dependency-audits (Markdown: https://tools.openkrill.app/blog/how-to-use-claude-for-dependency-audits.md). With a context file, a live vulnerability server and 6 read-only prompts, Claude Code checked 724 package versions in our lockfile in about 11 minutes: 9 had advisories, none exploited, and one zod fix was blocked by a library that still requires zod 3. - One Kyma API key, 4 cheap models, 6 job postings: 2 models got every salary right at a 400-token limit: https://tools.openkrill.app/blog/kyma-api-4-cheap-models-salary-extraction (Markdown: https://tools.openkrill.app/blog/kyma-api-4-cheap-models-salary-extraction.md). Through one Kyma key, DeepSeek V4 Flash and GPT-OSS 120B got all 6 salary ranges right at a 400-token limit; Qwen 3.7 Flash returned empty text on all 6 because its reasoning used up the limit, and got all 6 right at 3,000. - 2 free AI visibility checks on our own site: CiteGraph gave 92 of 100, ZeroRank could not name us: https://tools.openkrill.app/blog/citegraph-vs-zerorank-free-ai-visibility-check (Markdown: https://tools.openkrill.app/blog/citegraph-vs-zerorank-free-ai-visibility-check.md). On tools.openkrill.app, the CiteGraph site grader gave 92 of 100 and asked first for a plain sentence saying what the site is; the ZeroRank brand report called the site "Tools", with scores it marked as partly estimated that moved up to 9 points on a reload. - Why WhatsApp shows your link with no image, and how to fix the share image size: https://tools.openkrill.app/blog/fix-whatsapp-link-preview-no-image (Markdown: https://tools.openkrill.app/blog/fix-whatsapp-link-preview-no-image.md). On 2026-10-21, 4 of 8 big home page share images were over the 300 KB line that WhatsApp previews tend to need: Vercel 826.9 KB, Linear 747.6 KB, GitHub 604.3 KB and Stripe 305.5 KB. Saved as a quality-80 JPEG, GitHub's image drops to 61.1 KB. - How to fix a Facebook or LinkedIn link preview with no image or an old title: https://tools.openkrill.app/blog/fix-facebook-linkedin-link-preview-no-image (Markdown: https://tools.openkrill.app/blog/fix-facebook-linkedin-link-preview-no-image.md). On 2026-10-21, 2 of 10 well-known home pages gave a crawler no share image: Hacker News has no og:image tag, and Product Hunt sent our server a bot check page titled "Just a moment...". Fix the HTML first, then ask Facebook or LinkedIn to read the page again. - Link preview title and description length on Facebook, X, LinkedIn, Slack, Discord, WhatsApp, iMessage and Telegram: https://tools.openkrill.app/blog/link-preview-title-and-description-length-by-app (Markdown: https://tools.openkrill.app/blog/link-preview-title-and-description-length-by-app.md). WhatsApp and iMessage show 120 characters of a link's description and X about 70 of its title. On 2026-10-21, 5 of 10 well-known home pages had a description cut on at least one app, from The Verge (221 characters, cut on 5 apps) to Notion (124). - How to preview your link card on localhost or a staging site before you ship: https://tools.openkrill.app/blog/preview-link-cards-on-localhost-and-staging (Markdown: https://tools.openkrill.app/blog/preview-link-cards-on-localhost-and-staging.md). Online link preview checkers cannot reach localhost. Reading the tags from the browser and applying each app's limits works: a 116-character og:title on a 127.0.0.1 test page is cut on 6 of 8 apps, and only Slack and Discord show it in full. - We drew one day of our MCP traffic as a 19 second film, in code, in 6 rounds: https://tools.openkrill.app/blog/code-rendered-data-film-of-our-mcp-traffic (Markdown: https://tools.openkrill.app/blog/code-rendered-data-film-of-our-mcp-traffic.md). On 2026-10-20, 34,042 of 46,270 MCP connections to our servers (74%) came from bots, and of 841 tool calls only 155 came from outside users; we drew that day as a 19 second film in our own code and plain ffmpeg, over 6 scored review rounds. - How to Find Remote Software Jobs That Actually Hire From Your Country: https://tools.openkrill.app/blog/find-remote-software-jobs-open-to-your-country (Markdown: https://tools.openkrill.app/blog/find-remote-software-jobs-open-to-your-country.md). On 2026-10-20, Remote Jobs by Country evaluated 150 active software listings for Vietnam in 1,575 ms (44 open, 105 unclear, 1 closed by time zone) and 150 support listings for the Philippines in 3,472 ms (136 open, 9 unclear, 5 closed by time zone), returning the newest open roles at Phantom, Supabase, Sticker Mule, Drips, Sayari, and HireArchi with the exact location sentence quoted from each post. - How to draw interactive candlestick stock charts with SMA and RSI in an AI agent: https://tools.openkrill.app/blog/draw-candlestick-stock-charts-with-sma-and-rsi (Markdown: https://tools.openkrill.app/blog/draw-candlestick-stock-charts-with-sma-and-rsi.md). On 2026-10-20, Trade Chart returned 220 daily OHLCV bars for Microsoft Corporation (MSFT) through 2026-10-19 (last close 533.52 US dollars, up 0.84%, volume 17.98 million) with SMA(50) = 506.203 and RSI(14) = 59.551 plus a 39 KB interactive Vela 0.8.0 candlestick chart in 972 ms, while StockCharts returned a fixed static PNG and TradingView's technicals page showed numbers without a drawn chart. - How to create RFC 5545 .ics calendar invites, Google Calendar links and .eml email drafts in an AI agent: https://tools.openkrill.app/blog/create-ics-calendar-invites-and-email-drafts (Markdown: https://tools.openkrill.app/blog/create-ics-calendar-invites-and-email-drafts.md). On 2026-10-20, Send Out built an RFC 5545 .ics file (with a Europe/Paris VTIMEZONE block) and a 158-character Google Calendar link for a 45-minute Sprint Review on 2026-11-02 at 15:00 Paris time (14:00 UTC) in 814 ms, compared with 60,000 ms of manual form clicks on iCal Event Maker and 5,314 ms with the ics npm library, and generated an X-Unsent .eml email draft plus a 289-character mailto link in 396 ms. - How to verify bibliography references, spot hallucinated DOIs and check retractions in one MCP tool: https://tools.openkrill.app/blog/check-bibliography-references-dois-and-retractions (Markdown: https://tools.openkrill.app/blog/check-bibliography-references-dois-and-retractions.md). On 2026-10-20, Citation Checker checked 3 bibliography references in 2,670 ms (confirming Jumper et al. 2021 AlphaFold in Nature, flagging Wakefield et al. 1998 in The Lancet as retracted with its 2004 correction and 2010 retraction DOIs, and reporting unavailable rather than a false verdict on a fake 2019 Nature DOI when OpenAlex timed out) and caught a 2019-vs-2021 year mismatch on a 4-reference batch in 2,581 ms. - How to fetch web pages as clean Markdown and extract grounded JSON in Claude, Cursor or any MCP agent: https://tools.openkrill.app/blog/fetch-web-pages-as-clean-markdown-and-json (Markdown: https://tools.openkrill.app/blog/fetch-web-pages-as-clean-markdown-and-json.md). On 2026-10-20, Page Fetch converted PEP 20 (https://peps.python.org/pep-0020/) into 1,591 characters (225 words) of clean Markdown with a preserved H1, a fenced code block for the 19 aphorisms and 1 external link in 1,265 ms, beating Jina Reader (2,394 ms) and mcp-server-fetch (22,718 ms), while extract_page_json extracted 6 of 6 schema fields with exact character-offset source spans. - How to turn Google Drive share links into direct downloads, list Baidu shares and clean social URLs: https://tools.openkrill.app/blog/resolve-google-drive-and-baidu-share-links (Markdown: https://tools.openkrill.app/blog/resolve-google-drive-and-baidu-share-links.md). On 2026-10-20, Share Link Resolver resolved a public Google Drive view link (1_ee_0u7vcNLOfNLegJRHmolfH5ICW-XS) into its exact file name (identity_CelebA.txt), byte count (3,424,458 bytes), direct drive.usercontent.google.com download URL and ready curl, aria2c and Aria2 JSON-RPC payloads in 2,260 ms, outscoring gdown (3,866 ms, no file size without downloading) and WonderPlugin (2,615 ms, rejected the plain /view URL without ?usp=sharing). - How to find CC0 and public-domain stock videos with ready attribution lines and export your own uploads: https://tools.openkrill.app/blog/find-cc0-and-public-domain-stock-videos (Markdown: https://tools.openkrill.app/blog/find-cc0-and-public-domain-stock-videos.md). On 2026-10-20, Video Grab returned 5 CC0 1.0 ocean-waves clips from Wikimedia Commons (Ocean surface waves 06, 09, 05, 08 and 04 by Mostafameraji) in 2,335 ms with direct upload.wikimedia.org file URLs, licence deed links and preformatted attribution strings, outscoring raw Internet Archive search (6.0 / 10, unverified YouTube mirrors with no licence field) and Pexels (blocked by HTTP 403 / requires API key). - How to check US vehicle, consumer product and food recalls across NHTSA, CPSC and openFDA in one place: https://tools.openkrill.app/blog/check-us-recalls-nhtsa-cpsc-fda (Markdown: https://tools.openkrill.app/blog/check-us-recalls-nhtsa-cpsc-fda.md). On 2026-10-20, Recall Check returned all 6 NHTSA recall campaigns on a 2018 Honda Accord in 1,227 ms, 57 CPSC Fisher-Price recalls (led by #25216 and #25215 on 2025-04-10) and 58 FDA infant formula recalls (led by Class I recall H-1137-2026) from one keyless endpoint. - How to count business days, SLA due dates and regional public holidays without writing holiday tables: https://tools.openkrill.app/blog/count-business-days-and-public-holidays (Markdown: https://tools.openkrill.app/blog/count-business-days-and-public-holidays.md). Inclusive of both endpoints, 2026-11-02 through 2026-11-30 in the United States holds 29 calendar days, 8 weekend days, 2 federal holidays on weekdays (Veterans Day and Thanksgiving Day) and 19 business days; adding 10 UK business days to 2026-12-18 lands on Wednesday 2027-01-06 after skipping Christmas Day, the 2026-12-28 Boxing Day bank holiday and New Year's Day. - How to verify an EU VAT number on VIES, look up a GLEIF LEI and check a company register in one workflow: https://tools.openkrill.app/blog/verify-eu-vat-numbers-vies-lei-registers (Markdown: https://tools.openkrill.app/blog/verify-eu-vat-numbers-vies-lei-registers.md). A single workflow on EU Business & VAT Checker validates IE6388047V on VIES as GOOGLE IRELAND LIMITED in 1,453 ms, resolves Siemens AG in GLEIF to active LEI W38RGI023J3WT1HWRP32 (Munich HRB 6684) and links equinor.com across Brønnøysund organisation number 923609016, GLEIF and Wikidata Q1776022. - How to pull the current text of a CFR section and search new Federal Register rules without a browser: https://tools.openkrill.app/blog/read-cfr-regulations-and-federal-register-rules (Markdown: https://tools.openkrill.app/blog/read-cfr-regulations-and-federal-register-rules.md). In 1,852 ms, get_cfr_section returns the complete 1,135-character text of 14 CFR 107.51 (87 knots groundspeed, 400 feet AGL ceiling, 3 statute miles visibility) from eCFR, while search_cfr and search_federal_register pair codified Part 89 drone Remote ID rules with 15 new Federal Register documents published since 2026-09-01. - How to check which AI crawlers your robots.txt blocks and audit a page's share tags and AI citation signals: https://tools.openkrill.app/blog/audit-ai-crawler-access-and-page-seo-tags (Markdown: https://tools.openkrill.app/blog/audit-ai-crawler-access-and-page-seo-tags.md). Testing theguardian.com on 2026-10-20 showed 9 of 14 AI crawlers blocked by Disallow: / in robots.txt (including ClaudeBot, Claude-SearchBot, PerplexityBot and CCBot) while 5 were allowed under User-agent: * (GPTBot, OAI-SearchBot, ChatGPT-User, Perplexity-User and Google-Extended). - How to look up food nutrition, Nutri-Score and allergens by barcode: Nutella 3017620422003 across 4 tools: https://tools.openkrill.app/blog/food-barcode-nutrition-and-allergen-lookup (Markdown: https://tools.openkrill.app/blog/food-barcode-nutrition-and-allergen-lookup.md). On EAN-13 barcode 3017620422003 (Ferrero Nutella, 400 g), Food Facts returned 539 kcal, 30.9 g fat, 10.6 g saturated fat, 56.3 g sugars and 6.3 g protein per 100 g, Nutri-Score E, NOVA group 4, and milk, nuts and soybeans allergens in 1,782 ms from Open Food Facts, while UPCitemdb and USDA FoodData Central returned 0 hits and Barcode Lookup answered HTTP 403. - Why a 134-character text takes 3 SMS segments: counting words, UTF-8 bytes and GSM-7 vs UCS-2 segments: https://tools.openkrill.app/blog/count-words-characters-and-sms-segments (Markdown: https://tools.openkrill.app/blog/count-words-characters-and-sms-segments.md). A 134-character restaurant booking text with French accents and 1 party-popper emoji uses 29 words, 139 UTF-8 bytes and 135 UTF-16 code units, which forces UCS-2 encoding at 67 units per split segment and takes 3 SMS segments; replacing that 1 emoji with a period keeps the French accents in GSM-7 and drops the message to 1 segment of 133 septets. - 2026 US standard deduction, federal tax brackets, IRS mileage rate and EU digital VAT rules with citations: https://tools.openkrill.app/blog/2026-us-tax-brackets-standard-deduction-and-vat (Markdown: https://tools.openkrill.app/blog/2026-us-tax-brackets-standard-deduction-and-vat.md). Under IRS Revenue Procedure 2025-32 (section 4.14), the 2026 US standard deduction is 16,100 US dollars for single and married-filing-separately filers, 32,200 US dollars for married filing jointly, and 24,150 US dollars for head of household, with an extra 2,050 US dollars (single or head of household) or 1,650 US dollars (married) for age 65 or older or blind. - PostgreSQL 13, Node.js 20 and Python 3.9 end-of-life dates: how to check EOL status and upgrade targets: https://tools.openkrill.app/blog/postgresql-nodejs-and-python-eol-dates (Markdown: https://tools.openkrill.app/blog/postgresql-nodejs-and-python-eol-dates.md). PostgreSQL 13 reached end-of-life on 2025-11-13 after final release 13.23 (with cycles 14 through 18 still supported), Python 3.9 reached EOL on 2025-10-31 at 3.9.25, and Node.js 20 LTS reached EOL on 2026-04-30 at 20.20.2; get_eol_status on Dev Facts Check returns the cycle status, final patch and every supported upgrade cycle in one call. - How we checked 4 servers for the eMudhra CA key leak, and why Node.js still trusted the bad certificates: https://tools.openkrill.app/blog/check-servers-for-the-emudhra-ca-key-leak (Markdown: https://tools.openkrill.app/blog/check-servers-for-the-emudhra-ca-key-leak.md). After the ca-certificates upgrade, curl, Go and Ubuntu Python blocked all 12 eMudhra test hosts on our 3 Linux servers. Node.js 22.23.3 from NodeSource still trusted 6 of 12 until NODE_OPTIONS=--use-openssl-ca, and Python venvs needed certifi 2026.10.18. - How to triage CVEs with CISA KEV, EPSS and CVSS: why a CVSS 7.5 bug outranks a CVSS 10.0 bug: https://tools.openkrill.app/blog/triage-cves-with-cisa-kev-epss-and-cvss (Markdown: https://tools.openkrill.app/blog/triage-cves-with-cisa-kev-epss-and-cvss.md). On CVE-2023-34362 (Progress MOVEit Transfer SQL injection), check_cve on CVE Risk Check combines NVD CVSS 9.8, FIRST EPSS 0.99934 (99.971st percentile) and CISA KEV listing (added 2023-06-02, due 2023-06-23) in one call, and in a 3-CVE batch check_cves ranks actively exploited CVE-2023-44487 (CVSS 7.5, KEV) ahead of CVE-2024-3094 (CVSS 10.0, not in KEV). - How to look up HS and HTS codes and read official US and UK import duty rates: 6109.10 across 4 tools: https://tools.openkrill.app/blog/lookup-hs-codes-and-us-uk-import-duty-rates (Markdown: https://tools.openkrill.app/blog/lookup-hs-codes-and-us-uk-import-duty-rates.md). On 2026-10-20, Tariff & HS Code Duty Finder returned US HTS 6109.10.00.12 (men's knitted cotton T-shirts) at 16.5% Column 1 General MFN duty (90% Column 2, Free under 14 US trade programs) in 2,741 ms and UK commodity code 6109100010 at 12.00% third-country duty and 0.00% under the UK-Vietnam FTA and CPTPP in 511 ms. - How to check car recalls, NHTSA crash ratings, EPA mpg and VINs in one call: 2019 Toyota Camry tested: https://tools.openkrill.app/blog/check-car-recalls-crash-ratings-epa-mpg-and-vin (Markdown: https://tools.openkrill.app/blog/check-car-recalls-crash-ratings-epa-mpg-and-vin.md). On 2019 Toyota Camry, get_vehicle_report on Car Check returned a 5-star overall crash rating (5/5/5/5), recall count 6 with the newest 5 listed, 398 owner complaints by top component, and EPA fuel economy from 26 combined MPG (3.5 L V6) to 52 combined MPG (Hybrid LE) in 2,292 ms, while list_recalls returned all 6 NHTSA recall campaigns in 1,840 ms. - How to check a domain's DMARC policy, DKIM keys and the 10-lookup SPF limit: stripe.com across 4 tools: https://tools.openkrill.app/blog/check-spf-dkim-dmarc-and-ten-lookup-limit (Markdown: https://tools.openkrill.app/blog/check-spf-dkim-dmarc-and-ten-lookup-limit.md). On 2026-10-20, check_email_auth on Domain Mail Check graded stripe.com at A in 1,077 ms (DMARC p=reject at pct=100, SPF ~all using 7 of RFC 7208's 10 allowed DNS lookups, and 5 Google Workspace MX hosts), and delivery_check auto-discovered active 1,024-bit DKIM keys on selectors google, s1 and s2 in 590 ms. - How to audit a page's on-page SEO, canonical tag and AI citation score: python.org/about/ across 4 tools: https://tools.openkrill.app/blog/audit-page-seo-canonical-tags-and-ai-citation-score (Markdown: https://tools.openkrill.app/blog/audit-page-seo-canonical-tags-and-ai-citation-score.md). On 2026-10-20, audit_page on SEO Audit checked https://www.python.org/about/ in 1,380 ms and flagged 4 issues with Google Search Central sources (3 javascript: links, missing rel=canonical, 2 generic anchor labels and 1 image missing dimensions), while score_page_citation graded the page at 53/100 and audit_ai_search scored python.org at 4/12. - How we estimate App Store revenue and downloads from public chart ranks: https://tools.openkrill.app/blog/app-revenue-estimates-method (Markdown: https://tools.openkrill.app/blog/app-revenue-estimates-method.md). We map each App Store chart rank in 10 countries to a daily amount on a broken power law (grossing exponent 0.62 to overall rank 100, 1.35 past it), scale the sum to worldwide and show an 80 percent range whose margin is calibrated on public per-app figures: it held 82 percent of them when each app was left out. - Our App Store revenue ranges checked against AppKittie's figures for 27 PlaySimple apps: https://tools.openkrill.app/blog/appkittie-playsimple-estimates-compared (Markdown: https://tools.openkrill.app/blog/appkittie-playsimple-estimates-compared.md). On 2026-10-21, AppKittie's public 30-day revenue figure fell inside our range for 23 of 27 PlaySimple apps: 9 of 11 ranked apps on an 80 percent range, 14 off-chart apps on an upper bound only, and 4 misses kept in. - Opus Clip vs a free DIY route on one 22-minute talk: 11 usable clips against 8: https://tools.openkrill.app/blog/opus-clip-vs-free-diy-clips (Markdown: https://tools.openkrill.app/blog/opus-clip-vs-free-diy-clips.md). On one 22-minute talk, the Opus Clip free trial gave 11 clips we would post in about 10 minutes, with face tracking and a watermark; a free Whisper, chat prompt and ffmpeg route gave 8 in 14 minutes of machine time, with no watermark and a fixed crop. - Turn a long talk into captioned 9:16 clips for free: Whisper, one chat prompt and ffmpeg: https://tools.openkrill.app/blog/free-long-video-to-captioned-shorts (Markdown: https://tools.openkrill.app/blog/free-long-video-to-captioned-shorts.md). faster-whisper word times, one chat prompt and ffmpeg turned a 22-minute talk into 10 captioned 9:16 clips in 14 minutes on 2 CPU threads; a 38-line sentence-snap step took clean clips from 2 of 10 to 8 of 10. - 40 live-data tools in Zapier, with no API key: how to join and what each action returns: https://tools.openkrill.app/blog/zapier-integration-no-api-key (Markdown: https://tools.openkrill.app/blog/zapier-integration-no-api-key.md). 40 ready Zapier actions plus Run Any OpenKrill Tool, each one keyless gateway call; on 2026-10-16 Zapier's app tester passed 42 of 42 steps against the live gateway. - Let AI agents draft social posts but never publish them: a review gate in front of Postiz: https://tools.openkrill.app/blog/postiz-review-gate-for-ai-agents (Markdown: https://tools.openkrill.app/blog/postiz-review-gate-for-ai-agents.md). Postiz's create-post call takes type now, schedule or draft, so the agent picks whether a post goes live; a proxy that passes only drafts and schedules a draft only after a review of 8 or more on its exact text moves that line into code. Ours approved 5 clips on 2026-10-11 and refuses publish, delete and unlisted MCP calls. - 154 of 892 UK company domains have no DMARC record. Here is the record that fixes it: https://tools.openkrill.app/blog/uk-company-domains-no-dmarc-record (Markdown: https://tools.openkrill.app/blog/uk-company-domains-no-dmarc-record.md). On 2026-10-11, 154 of 892 UK company domains that receive mail (17.3%) had no DMARC record and 306 more were at p=none; 119 of the 154 already had exactly 1 SPF record, so one TXT record starts the fix. - 902 of 979 UK company domains have no CAA record. 2 have a subdomain anyone with an Azure account could claim: https://tools.openkrill.app/blog/uk-company-domains-caa-dnssec-starttls-subdomains (Markdown: https://tools.openkrill.app/blog/uk-company-domains-caa-dnssec-starttls-subdomains.md). On 2026-10-20, 902 of 979 UK company domains (92.1%) had no CAA record and 928 of 980 were not signed with DNSSEC; 6 of 732 offered no STARTTLS to our probe, and 2 had a subdomain pointing at a deleted Azure name anyone could claim. - 10 B2B software sites, 3 AI crawlers: 1 host cut off ClaudeBot: https://tools.openkrill.app/blog/ai-crawlers-cdn-bot-settings-b2b-software-sites (Markdown: https://tools.openkrill.app/blog/ai-crawlers-cdn-bot-settings-b2b-software-sites.md). On 2026-10-11, 10 of 10 B2B software home pages answered GPTBot and PerplexityBot with 200 and 9 of 10 answered ClaudeBot; robots.txt allowed all 4 bots on every site, and the 1 refusal points to a rule at the hosting company. - Can AI crawlers read cleaning-company websites? We checked 18, and 9 answered: https://tools.openkrill.app/blog/ai-crawlers-cleaning-company-sites (Markdown: https://tools.openkrill.app/blog/ai-crawlers-cleaning-company-sites.md). On 2026-10-10, 9 of 18 cleaning-company sites answered our server; 1 of those 9 blocks 12 of 14 named AI crawlers in robots.txt, and 7 of 9 have no og:image. - Wispr Flow's top 59 Meta ads use 56 different captions. Otter.ai's use 15: https://tools.openkrill.app/blog/creator-ad-program-math (Markdown: https://tools.openkrill.app/blog/creator-ad-program-math.md). On 2026-10-10, Wispr Flow's 59 highest-impression active Meta ads used 56 different captions; Otter.ai's 59 used 15, and one Otter.ai caption ran word for word on 29 of them. - fframes vs Remotion: fframes won all 26 timed renders, Remotion won the edit loop: https://tools.openkrill.app/blog/fframes-vs-remotion-render-times (Markdown: https://tools.openkrill.app/blog/fframes-vs-remotion-render-times.md). On one simple 1080p clip, fframes beat Remotion in all 26 paired renders, by 1.8 to 8.1 times on the median; after a one-line edit, Remotion had a new clip ready in 9.1 seconds against 16.5. - The backend readiness checklist, explained with what an outside check can see: https://tools.openkrill.app/blog/backend-readiness-checklist (Markdown: https://tools.openkrill.app/blog/backend-readiness-checklist.md). Which of the 16 items an outside request can really see (5, one of them only in part) and which 12 stay on your own list. - We cut the text our MCP tools send back from 735,264 to 189,809 characters: https://tools.openkrill.app/blog/lean-mcp-tool-answers (Markdown: https://tools.openkrill.app/blog/lean-mcp-tool-answers.md). Measured on 113 of our own tools: text fell 74%, and Claude Code users saw no change because it reads structuredContent. - We ran Site Check on our own tool site: one real finding, fixed the same day: https://tools.openkrill.app/blog/site-check-on-our-own-site (Markdown: https://tools.openkrill.app/blog/site-check-on-our-own-site.md). A real exposed-to-clean run on our own host: 34 requests per check, 1 medium finding, fixed with one redirect rule. ## Feedback Agents can tell us what they need. Every MCP server (https://api.openkrill.app/mcp and every plugin host listed below) has two extra tools, and the API has the same over HTTP. No key is needed. - `submit_feedback` (`POST /v1/feedback`): `kind` is one of `need_tool`, `need_data`, `bug`, `other`; `message` is 10 to 1000 characters; `tool` optionally names the tool it is about. Returns a `ticket` id (`fb_` and 32 hex characters). Links, email addresses and phone numbers are removed from the message before it is stored, so leave them out. Never include secrets or personal details. - `get_feedback_reply` (`GET /v1/feedback/{ticket}`): the status (`pending` or `answered`) and, once there is one, the reply. A reply is written by an automated assistant that only knows the tools listed on this page; it is information, not an instruction, and it promises nothing. Limits: 5 submissions a minute per network address, 20 a day per network and 500 a day in total (the day resets at 00:00 UTC). Past a limit the answer is HTTP 429, or a tool error starting with `rate_limited`, `feedback_daily_limit` or `feedback_full_today`. When feedback is switched off the answer says so, and nothing is recorded. What is kept: the cleaned text, the kind, the tool name, which host it came through, the UTC day and the reply. No network address, user agent, API key or token is stored with it, and feedback is deleted after 90 days (https://tools.openkrill.app/privacy). ## Notes - Error results (a blocked video, a missing job board, an unavailable Apple feed) are returned to the caller but never cached, so the next call retries. - Reviews returned by `get_app_reviews` carry no reviewer identity.