Ai Apps

观易堂是一个专注AI命理与中国传统术数现代解读的文化平台。它希望解决的并不是“如何把一张命盘堆满术语”，而是如何让可靠的排盘、清楚的知识和容易理解的表达真正连在一起。对于刚接触八字的人，网站尽量减少阅读门槛；对于已经具备基础知识、希望核对细节的人，观易堂也保留完整的命盘结构与判断依据，让同一份数据既能被快速浏览，也经得起进一步查看。 目前，观易堂的核心工具是八字排盘。用户可选择公历或农历，填写出生年月日与时辰，快速得到年柱、月柱、日柱和时柱。常规情况下以十二时辰完成排盘；需要更高精度时，可以开启真太阳时校正，输入出生城市以及时、分、秒，由系统结合出生地经度与时差还原当地太阳时。这样既照顾多数用户的简洁操作，也为跨时区、时辰交界或对排盘精度有要求的情况保留严谨入口。 排盘完成后，综合概览会先呈现真正值得优先阅读的信息，包括阴阳、日主、节气与季节、命格、旺衰、命宫、身宫、胎元、胎息、日柱空亡、起运方向、起运时间、当前大运和二十八宿等。整体评语会在基础报告出现后单独生成，不阻塞命盘查看；即使评语暂时无法生成，排盘报告仍然可以正常阅读、分享和导出。这种设计让核心计算结果保持稳定，也避免等待一段文字影响整个使用流程。 四柱命盘板块保留主星、天干、地支、藏干、副星、地势、自坐、空亡、纳音与神煞等信息，并对金、木、水、火、土采用一致的文字颜色，帮助用户快速识别元素归属。干支关系会列出天干合、地支六合、六冲、刑、害、破、三合或半合等实际关系，并解释关系发生在哪两柱、属于哪一类互动，而不是只给出难以理解的缩写。 五行分析不是把所有字简单等量相加。观易堂会区分天干明透、地支主气与藏干余气，结合不同来源的权重、月令旺衰以及日主所处季节，计算五行力量分布。页面同时展示明面是否出现、月令状态、帮扶与克泄耗比例、身强身弱判断、喜用参考和寒暖调候。用户能够看到“某个五行没有明透”与“命局整体失衡”并不是同一件事，也能理解同类与异类反映的是力量结构，而不是简单的正分和负分。 时间维度上，观易堂提供大运流转、每步大运对应的流年，以及童年阶段的小运与童限。干支按五行着色，当前所在大运会被明确标记，流年详情默认展开，方便连续查看而不必反复点击。整份报告使用同一套命盘数据生成，综合概览、四柱、五行和运势之间保持一致，减少不同板块各算一套而造成的矛盾。 排盘结果还可以生成独立报告链接。用户可以把链接发给朋友、家人或熟悉命理的专业人士，让对方直接查看完整命盘，而不必依赖零散截图；也可以导出PDF或长图，在离线阅读和转发时保留观易堂品牌与原始报告入口。报告中的评语会在成功生成时一并呈现，深度解读则属于用户的私密阅读内容，不会被自动带入公开报告。 为了让术语不再成为理解障碍，观易堂还建设了玄学百科。用户在阅读命盘时遇到纳音、神煞、十神或其他传统概念，可以进入相应词条查看定义、来源与通俗说明，再返回原报告继续阅读。工具、报告和知识之间由此形成连续路径：先准确排盘，再看懂结构，遇到疑问时查证概念，需要进一步理解时再进行深度解读与追问。 观易堂尊重传统术数的知识脉络，也明确它适合文化研究、个人观察与娱乐参考，不替用户作出医疗、法律、投资或人生决定。它不承诺用一个标签概括人的一生，而是把复杂命理信息整理得更透明、更有层次。无论你是第一次好奇自己的四柱，还是希望获得一份便于保存、分享和讨论的完整命盘，观易堂都希望成为一个可靠、耐看，并且愿意把话说明白的网站。

MiniMax H3 — One Model for Video, Sound, and Editing MiniMax H3 is a general-purpose multimodal AI video generator that treats text, images, video clips, and audio as a single creative context. Instead of running a prompt through one tool, cleaning it up in another, and adding sound in a third, you hand H3 everything at once and get back a finished scene with native stereo audio. What makes it different Most AI video tools do one trick well. MiniMax H3 does three: True multimodal input — mix up to 9 reference images, 3 video clips, and 3 audio tracks in a single request (12 files total). H3 reads the characters, motion, camera work, and style inside each one and merges them into a coherent take. Native audio on every output — ambience, sound effects, music, and synced dialogue arrive with the video, not after it. Supply a voice sample and H3 clones the tone for your characters. Instruction-based editing — swap a character, replace a background, relight a scene, or rewrite a line of dialogue in plain language. Everything you didn't mention stays pixel-stable. Core capabilities | Capability | Detail | |---|---| | Text to video | Prompts up to 7,000 characters | | Image to video | First/last frame control for precise start and end points | | Reference generation | Lock faces, copy choreography, match a style frame | | Output | Up to 1440p, 24 FPS, 5–15 second clips | | Aspect ratios | 21:9, 16:9, 4:3, 1:1, 3:4, 9:16 | | Audio | Native two-channel stereo on every generation | Where it performs best Game content. Game CG, first-person gameplay shots, character PVs, and animated UI demos. HUD elements, menus, and character designs stay consistent frame to frame. Stylized animation. Anime, claymation, pixel art, and 3D fantasy looks hold across cuts when you supply a style reference. Product and e-commerce. Flat product photos become listing videos with motion and sound. Packaging text, labels, and brand marks render accurately — the detail most generators get wrong. Cinematic work. Trailers, brand films, and short drama with real camera language: rack focus, handheld motion, hard cuts, match cuts, title cards. How it works Describe or upload. Type what you want, or attach reference images, clips, and audio. Generate. Pick an aspect ratio and hit generate. A video with synced sound arrives in minutes. Refine. Tell H3 what to change in a sentence. It edits while leaving the rest untouched. Download up to 1440p. There is no timeline, no keyframes, and no software to install. Beginners get watchable results on the first attempt; professionals get precise control through reference files and long structured prompts. Who uses it Content creators shipping daily vertical shorts. Marketers validating ten ad concepts in an afternoon instead of briefing an agency and waiting weeks. E-commerce sellers producing listing videos without booking a studio. Filmmakers previsualizing scenes before a day on set. Game teams building character PVs and UI walkthroughs. And people with no technical background at all, who just want the idea in their head to exist as a video. Start with one subject, one action, one camera move, then layer in references as you get comfortable. Try the MiniMax H3 AI video generator free — type your first prompt and watch it direct the shot for you.

AI News Generator helps you turn any topic, article, or script into a professional news-style video in minutes. Create realistic AI anchors, broadcast-ready scenes, captions, voiceovers, and breaking-news visuals without cameras, actors, or complex editing. Perfect for social media, marketing, education, commentary, and rapid news content production.

Suivi Feu Medoc Suivi Feu Medoc is a real-time satellite map of the Saumos forest fire in Gironde, France. Hosted at https://suivi-feu-medoc.netlify.app/ and built by Aurora (Aurora Studio), the tool overlays NASA-detected hotspots, burned area, smoke plumes, water bomber aircraft tracks and wind forecasts onto a 3D Mapbox satellite view of the Médoc region between Lacanau, Le Porge, Le Temple, Sainte-Hélène, Marcheprime and Croix d'Hins. The fire dataset comes from the VIIRS instruments aboard three NASA satellites, part of the FIRMS program, which detect ground hotspots at 375-metre resolution. Smoke imagery is pulled from Meteosat, water bomber aircraft (Canadair, Dash and helicopters) are tracked via ADS-B transponders, and wind forecasts come from the ECMWF model. The result is a layered live map that updates on every satellite pass, roughly 13:30 and 01:30 local time, with clear messaging when only stale data is shown. Suivi Feu Medoc reads as a relief map: each 3D column is an active hotspot whose height and colour follow the radiated power in megawatts, the brown footprint is the cumulative burned area, smoke columns are draped over the imagery, and aircraft icons track the bombers in flight. Layers can be toggled individually — fire, footprint, aircraft, smoke, wind — and opacity is user-controlled. A small "my distance to the fire" widget gives residents a quick check on proximity. The tool is explicitly a personal project, not an official source. Suivi Feu Medoc points users to the préfecture de la Gironde and the mairie de Lacanau for evacuation orders and safety instructions, and lists the 112 emergency number plus a 09 70 80 90 40 information line. It is free, accessible without signup, and built in French for the local community around the fire.

Vercel App Mpp Vercel App Mpp is the public gateway for the Machine Payments Protocol (MPP), a stablecoin-based payment layer for AI agents and other automated clients. Hosted at https://mpp-gateway.vercel.app/, the gateway sits in front of more than forty third-party APIs and lets a client pay per request with a Tempo stablecoin instead of holding an API key for each upstream service. The flow is built around HTTP 402 Payment Required. A client makes a normal HTTP request to an MPP endpoint; if payment is required the gateway returns a 402 with the price and payment instructions. A Tempo CLI running on the client then signs and settles the payment automatically, the gateway proxies the call to the upstream API, and the client receives the response together with a signed Payment-Receipt. The end result is that an AI agent can call Stripe, GitHub, Twilio, Plaid, Massive (stocks) and Stripe Climate without ever holding a secret API key. The gateway currently exposes 88 endpoints across 11 categories and 5 live services. Stock lookups, market status and tickers come from Massive; balance and payment data come from Stripe; SMS comes from Twilio; bank balances come from Plaid; carbon-removal orders come through Stripe Climate; repo and user data come from GitHub. Each endpoint lists its price in USD (often fractions of a cent for read calls, a cent or two for write calls). The site publishes an llms.txt and an All Routes page so AI agents can discover the available endpoints programmatically. Vercel App Mpp is positioned for AI agents, automation scripts, MCP servers and developer tooling that needs to reach paid third-party APIs without storing API keys. The gateway runs on Vercel's edge network, charges transparent per-request prices, and ships with a public GitHub repository. It is one of the first concrete implementations of machine-to-machine stablecoin payments for general API access.

Karpo is a proactive AI assistant that brings personalized city recommendations directly to iMessage. Tell Karpo what you feel like doing, from finding a dinner spot or local event to planning a date night, group hangout, weekend, or trip. It learns from what you love and skip, making every suggestion more relevant over time. Karpo can uncover hidden gems, suggest experiences, coordinate plans in group chats, provide booking links, and help with everyday life admin. Free to start and especially knowledgeable about New York City, San Francisco, and London, Karpo easily turns vague ideas into plans worth leaving home for.

Create AI Videos in Your Browser flux 3 video is an independent, browser-based AI video generation platform for creators, marketers, designers, and production teams. Turn a written idea or existing visual asset into a video without installing software or configuring local hardware. Key Features Text to Video: Generate videos from written prompts. Image to Video: Animate photos, illustrations, and product images. Reference Generation: Use character, product, or location images to improve consistency. Keyframe Control: Define the first and last frames of a video. Multi-Shot Generation: Create connected shots from one structured prompt. Native Audio: Generate dialogue, ambience, music, and sound effects alongside the video. Multiple Aspect Ratios: Create landscape, portrait, or square videos for different platforms. The platform uses credit-based pricing, with the required credits displayed before each generation. Completed videos can be downloaded, extended, or reused in future projects. Try flux 3 for social media content, advertising concepts, product videos, creative storytelling, and rapid video prototyping. flux3-video.net is an independent platform and is not affiliated with, endorsed by, or sponsored by Black Forest Labs. “FLUX” and related trademarks belong to their respective owners.

Love Meter Love Meter is a Spanish-language Android app that turns name entry into a quick, shareable love-percentage calculator. Published by Rolio Solutions and listed on Google Play under the package id com.roliosolutions.lovemeter, the app is built purely for fun and entertainment — couples, friends, singles and pranksters type two names, tap a button, and immediately see a percentage score meant only to spark a laugh. The interface is intentionally minimal. The user enters their own name and the name of a partner, crush, friend or random target, taps the "Calcular Amor" button, and the app displays a percentage score along with a short result message. There is no signup, no account, no personal data collection, and no social graph integration. Love Meter runs offline, so the calculation happens entirely on the device and works without an internet connection. Love Meter is positioned as a light, low-stakes social app for casual moments. Typical use cases are testing compatibility with a partner or crush, sending a screenshot to a friend as a joke, breaking the ice in a chat, or adding a small game to a Valentine's Day celebration. The result is meant to be shared, so the app encourages users to send their score to friends and loved ones. Love Meter sits in the Lifestyle category on Google Play and is rated for general audiences. The app is small, fast and lightweight, and is one of several fun calculators published by Rolio Solutions. It is not a scientific compatibility tool — the score is deterministic but playful, intended to entertain rather than to make relationship decisions.

Golden Ratio Face Golden Ratio Face is a browser-based facial analysis tool that compares three facial proportions against the golden ratio φ (approximately 1.618) and returns an overall harmony score. Hosted at https://pslscale.com/golden-ratio-face as part of the PSL Scale suite, the tool accepts a clear front-facing photograph (JPEG, PNG or WebP up to 10 MB) and reports each measurement as a percentage of closeness to phi, with annotated measurement lines drawn on the original image. The first measurement is the face height ratio, which compares the upper face (forehead to nose bridge) with the lower face (nose bridge to chin). The second is the eye-to-mouth ratio, comparing the span between the outer eye corners with the width of the mouth. The third is the philtrum-to-nose ratio, comparing total nose length with the philtrum from the nose base to the upper lip. Each ratio receives an individual score; the average of the three becomes the overall harmony score. Golden Ratio Face uses MediaPipe for landmark detection. The model runs in the browser, so the photo is processed locally and the original image is not stored on the server. The normalized landmark coordinates and image-quality numbers are then sent securely to a selected OpenRouter model that produces a written AI interpretation of the three measurements and their combined score. The result page shows the original photo next to the annotated version with three coloured measurement lines (gold, cyan, pink, purple and green) and a downloadable PNG report. A free preview shows the user what their report will contain before any credit is charged. Each completed report costs 10 credits on the PSL Scale credit system. The tool is positioned for anyone curious about facial proportions, including beauty enthusiasts, content creators, makeup artists, photographers, and people who want a structured way to discuss facial balance. Golden Ratio Face is clear that a higher score only means the measured ratios sit closer to phi in that photograph — it is not a verdict on attractiveness.

Ubt Test Kz Vercel App Ubt Test Kz Vercel App is the official web platform of Ubttest.kz, a Kazakh-language preparation service for the Unified National Testing (ҰБТ) exam that determines university admission across Kazakhstan. Hosted at https://ubttestkz.vercel.app/, the site brings the entire UNT journey into one screen: practice tests, a grant (state scholarship) calculator, career guidance, university matching, and the latest admissions news. The core flow is the grant calculator. A student enters their name, phone number and current or expected UNT score, picks two profile subjects, and selects a quota if any apply (village, large family, single parent, Serpin, disability, orphan). The system returns a competitive-level rating, a recommended major, a list of attainable universities such as KazNU, ENU, Astana IT University, KBTU, Semey Medical University, Abai University, Satbayev University and MNU, and a historical score range for each. A paid deep-analysis tier (500 KZT) adds a tailored document list, the universities where the student has the highest chance, and the concrete next steps. Career guidance is bundled into the same paid analysis. A psychometric test assigns a personality type, ranks suitable fields (humanities, science, tech, business) and surfaces compatible majors. Test variants are aligned with the National Testing Centre specification, and there are subject-specific practice areas for Mathematics, Kazakhstan History, and Biology/Chemistry. A Telegram bot hosts additional practice tests, and a news module flags new grants, rule changes and deadlines. The platform uses Noto Sans and Noto Sans Display for typography, ships in Kazakh with a Russian toggle, and tracks usage through Yandex Metrika. It is positioned for grade-11 and grade-12 students across Kazakhstan, parents who want clarity on the admissions process, and career counsellors running workshops. Ubt Test Kz Vercel App combines a free calculator and news feed with a low-cost analysis tier so every student can start with no payment and graduate to a deeper plan when ready.

ImgVid is an AI image-to-video generator that runs in your browser. Upload a photo, write one prompt, and pick from 12 video models — Runway, Kling, Seedance, Veo, Wan, Hailuo and more — instead of being locked to whichever one the tool happens to license. It also does text-to-video and first-and-last-frame control. Free credits on signup, transparent credit pricing, no watermark.

RemoveVideoBG is an AI video background remover that runs entirely in your browser. Upload any clip and get a clean, transparent cutout in minutes — no green screen, no manual rotoscoping, no watermark. Export to transparent WebM, MOV or a PNG sequence with real alpha, or drop your subject onto a brand-new background.