Complete case descriptions without leaving the website.
Restarting the board resets the timer. It counts the ESP32 power supply time, not user activity: if USB remains on during sleep computer, counting continues. After 99:59:59 the clock goes to 00. Animation: 16 single-color frames, 1920 ms cycle, 16 KiB in flash. White numbers on a black background at the bottom of the screen. Wi-Fi is not used. oled-preview.gif shows the prepared background without a timer.
stack Front Next.js 16 (App Router) + React 19, TypeScript, Tailwind v4, lucide-react. Login via Privy (@privy-io/react-auth v3) - Solana wallet as the only authorization method. Transactions are collected on @solana/kit v7 + web3.js, USDC transfers via @solana-program/token. Back NestJS 11 on Node 24, TypeScript. Prisma 7 → PostgreSQL 16, the scheme is divided by domain into 12 files (orders, balance, cashback, levels, referrals, refund-requests, notifications, admin-settings), 13 migrations. Input validation - class-validator on DTO. Tests - Jest. Redis (ioredis) keeps quotes with TTL - the price lives exactly until the timer expires, and is recalculated when signed. @nestjs/schedule: hourly crowns unlock cashback, interval jobs pull order statuses. Webhooks are deduplicated through the table of processed events - the repetition will not happen twice. Solana Payment is verified by the signature on the back: getTransaction with retrays and two RPCs (main + fallback), separate branches for native SOL and SPL transfers. Token rate
replenishment of Steam and Gift cards for crypto You pay from your wallet in SOL or USDC - the code arrives in a couple of minutes. Without banks and intermediaries. • Steam wallet: USD, RUB, KZT, UAH • Gift cards, game keys, service replenishments • Cashback 15% in $QINTO from each order • Referral: 1% from friends’ deposits, also in a token • Internal balance: payment in one tap, private key and withdrawal are always yours • Bot @Qinto_bot - all the flow in the cart Replenishment failed? We will automatically select gift cards for the same amount, and we will bear the difference up to $1. This didn’t work either - the money would be returned to the wallet on its own, without tickets. Solana, payment in SOL/USDC with reconciliation of each transaction on the chain. $QINTO - Token-2022. Course from Jupiter, quote 3 minutes. NestJS, PostgreSQL, Next.js, Privy, grammY, keys for AES-256-GCM. The principle of the monetary part: first entry into the ledger, then transfer. Payments are not lost, each order pays exactly once.
A fully automatic system that catches undervalued skins on mrkt.land and compares them with buff163 prices in real time. Architecture - 4 accounts, each with its own role 3 accounts on mrkt.land + 1 on buff163, working in parallel in one process: • Account feed - reads the “New First” stream, catches newly exhibited lots • Account catalog - goes through the entire catalog by price, catches overpriced lots (they are not in the “new” ones, but they often hide benefits) • Buyer account – holds the balance and makes a purchase • buff163 — background price sync: ask (showcase) and bid (buy-order) Separation by accounts gives each one its own IP and its own session - the feed and directory are polled simultaneously, without being limited by the general rate-limit. Dolphin Anty (anti-detect) Each account lives in a separate Dolphin profile with its own resident proxy. The key point is IP consistency: from which address the account logged in, all requests come from the same. For the site, it looks like a regular live browser, and not like a bot from one server. Proxies are protected by profiles, so nothing breaks when transferred to hosting - traffic still goes through the “native” IP accounts. How to parse (not DOM, but API) mrkt.land is Next.js on top of a private JSON-API. We work directly with the API (endpoints are pulled from the JS bundle), rather than parsing HTML. This is an order of magnitude faster - profitable lots are sorted out in a fraction of a second. An important nuance: the feed is a stream of events, not a list of products. There are 4 types: listing (exposed), change_price (overpriced - still on sale), unlisting (removed), sale (already sold). We buy only the first two + Validated status. We cut off the rest so as not to chase after lots that no longer exist. Logic of price and benefit For each skin, we count two landmarks at once from buff163: • ask (showcase price) - the main threshold: we take it if mrkt is at least 15–25% cheaper • bid (buy-order, instant sale) - control: usually +3–8%, shows the real output “here and now” • bid (buy-order, instant sale) - control: usually +3–8%, shows the real output “here and now” Both numbers are written for each skin found, so you can see both the “paper” benefit and the real one. Filters and fuses • We exclude StatTrak, Souvenir, Well-Worn; skins from $20 • Exact name matching (no fuzzy - mistakes cost money) • Stop at rotten buff prices, outdated rate, thin market (<3 lots), absurd profit (matching error), exceeding limits per trade/day • Protection against double purchases and “slippage” (purchase at exactly the price at which profit was calculated) Management • Web panel: live log, counters, start/stop cycles on the fly, switching DRY_RUN live buyback, setting profit thresholds and limits directly from the browser • Notifications in Telegram for each purchase - with a picture, float, price, profit and link to the item • Hosting under systemd, access via IP About speed Everything is tied to being the first to receive and evaluate data. Warm keep-alive connections, polling the tape once a second, instant reading of prices from the local cache. The real ceiling is physics: some of the truly profitable lots are taken by co-locking bots in ~100 ms, but the system honestly takes everything it manages to reach.
It was an AI agent capable of independently improving itself. He used Anthropic providers to generate code in a remote sandbox, where he automatically: generated new code; ran tests; created Change Log; posted updates to X; updated the page with the latest changes. Project stack: Next.js - frontend; NestJS - backend; PostgreSQL is the main database; Redis - for caching and stable system operation. The agent generated MJS code, independently extended its rules, and even created new tools. For example: Solana Wallet Checker; various utilities for automation; and many other useful tools. Unfortunately, the project didn't end the way I wanted... He fell victim to a rug pull, and all the work he did just disappeared.
stream assistant that just works I put together a native application for streamers: one installer, a couple of clicks “Next” - and you have chat aggregation, alerts, an overlay for OBS, chat voiceover and connection to OBS without a single manual configuration of Browser Source. No Python, no “first install an interpreter and 20 libraries” - everything is in C++, it works right out of the box. What's inside: — Chat from Twitch and YouTube in one overlay, with alerts for subscriptions/donations/raids — Chat voice-over (TTS) — with optional voice change via RVC (you can even answer in Ayaka’s voice in chat) — Volume auto-ducking — during voice acting, other applications are muted themselves — Polls in chat - viewers vote !1/!2, the result is immediately on the overlay — Music by points - viewers accumulate points for activity in the chat and order tracks from YouTube/SoundCloud with the !song team — Automatic connection to OBS — the application itself creates the necessary sources, you don’t need to add anything manually Everything is lightweight and modular: TTS and voice change are separate optional add-ons, installed with a button only if really needed. Download and try: https://github.com/PRISSET/StreamSoftNative/releases/latest
It seems not bad, still the same Next.js
Essentially, this is a real Telegram combine for warming up accounts, working with traffic, finding an audience and many other tasks. The project is currently in the testing stage and does not yet reach a full-fledged MVP, so your feedback is especially important to us. At the moment, account warming is available free of charge for all users. If you are interested in tools for working with Telegram, come, test and evaluate the project: link For each bug found, 1 will be allocated to your balance, there are many bugs
I made a loader, a flawed one in C++ downloads dependencies, also dlls from Git via raw and injects the cheat.
You can look at the VPN test project - @xsmoke_bot For now, this is just a test and testing of some ideas, but it is quite possible that you will soon see this project in production
Project information: --- Frontend - Next.js + TypeScript What has been implemented: — UI from scratch: landing page, custom dashboard, admin panel — Dark minimalistic style with tech aesthetics — Manrope + Orbitron fonts — Tailwind CSS + CSS Modules — Interactive SVG map of server locations (Sweden, Netherlands, Germany, Poland) — CSS animations: reveal via Intersection Observer, ticker, shimmers, confetti — Adaptive from mobile to 4K — Server configurator with live price calculation — Dashboard: servers, payment, profile, tickets, notifications, referral — Level and XP system --- Backend - Nest.js + PostgreSQL + Prisma Architecture: — Modular structure: Auth, Users, Servers, Payments, Tickets, Notifications, Referrals, Admin — REST API with validation (class-validator) — JWT authentication via Passport.js, separate for users and admins — Prisma ORM with multi-file schema - RBAC with granular permissions What has been implemented: — Integration with VMmanager (VPS provisioning, metrics, management) — Telegram Bot: account linking, notification delivery — Multichannel notifications: in-app, email, Telegram — 4 payment providers: YooKassa, Fride, Heleket, CryptoBot — Support for SBP, RF/EU cards, crypto — Dynamic USD/RUB rate — IP ban at the middleware level Admin panel: — Dashboard with real-time cluster metrics — Management of users, servers, tariffs — Registration and payment schedules — Audit logs — System settings --- DevOps — Push in Git — autodeploy — Migrations via Prisma — Automatic restart of services --- Stack: Next.js / Nest.js / PostgreSQL / Prisma / Tailwind CSS / Telegram API / VMmanager / YooKassa / JWT / Passport.js link to the project - https://nodehost.cc/
Project information: Next.js What has been implemented: — Completely designed from scratch UI — Minimalistic premium style (b/w, large typography, emphasis on the brand) — Custom fonts — GSAP animations (smooth appearances, dynamic transitions, working with typography) — Adaptive layout — Loading and performance optimization Visually, there is an emphasis on cleanliness, air and large headlines that create the feeling of a digital brand. Nest.js + PostgreSQL Architecture: — Structured by modules — Clean code organization - REST API — Working with PostgreSQL through migrations Additionally implemented: — Integration with Telegram API (creation and management of sweepstakes) — Admin panel: • adding products • category management • management of managers • content editing Autodeploy configured: - push to Git - error checking — performing database migrations - restarting services The project is updated automatically without manual intervention. : Next.js / Nest.js / PostgreSQL / GSAP / Telegram API
