Bookmarks

Cool Websites — June 27, 2026

Security & Privacy

SimpleX Chat

The only messenger with no user IDs whatsoever — not even an anonymous random number. Instead, each conversation uses a throwaway one-way queue address; servers route messages but can never correlate who talked to whom or build a social graph. Fully open source, end-to-end encrypted, and available on iOS, Android, and desktop. The strongest metadata privacy of any mainstream messenger.

BrowserLeaks

A comprehensive privacy audit suite that reveals exactly how trackers can identify you despite a VPN or privacy browser. Tests cover WebRTC IP leaks, Canvas/WebGL fingerprinting, installed fonts, geolocation, battery API exposure, and a dozen more attack vectors — all explained clearly so you understand what you're fixing. No data collected; everything runs in your browser.

OnionShare

Open-source peer-to-peer file sharing over Tor — your own machine becomes the web server, so no cloud intermediary ever handles your files. Supports anonymous file sending, receiving, chat rooms, and even hosting a temporary website, all behind a `.onion` address. Pre-installed on Tails, Qubes, and Parrot OS; used by journalists, whistleblowers, and anyone who needs to share sensitive material without a paper trail.

OSINT & Investigation

LeakIX

Unlike traditional port scanners that just report "port 9200 open," LeakIX actually reads what's exposed — open Elasticsearch clusters with document counts, publicly writable MongoDB collections, leaked `.env` files with database credentials, exposed Git repos, and unauthenticated admin panels. If you need to audit your own attack surface or understand what real exposure looks like in the wild, it's one of the most informative tools available.

GreyNoise

A global deception network that catalogues the constant background radiation of internet scanning — all the automated bots, vulnerability scanners, and mass exploitation tools probing every IP address every few minutes. GreyNoise tells you "this IP is just a benign Shodan crawler" vs. "this IP is actively exploiting CVE-2025-XXXX right now," cutting SOC alert noise by 60–70%. Free community tier gives access to the live feed; used by Fortune 1000 security teams.

BuiltWith

Enter any domain and see every technology it uses: CDN, analytics, ad networks, CMS, ecommerce platform, payment processors, A/B testing tools, and more — with historical snapshots so you can see when they switched. Invaluable for competitive intelligence ("what stack does every $10M SaaS run on?"), sales targeting, and tracking technology adoption trends across the web. The free version covers most lookups; Pro unlocks bulk CSV exports and CRM integrations.

IT & Sysadmin Tools

MXToolbox

The Swiss Army knife for email infrastructure diagnostics: look up MX records, run SPF/DKIM/DMARC validation, check your mail server against 105+ DNS blacklists, test SMTP connectivity, reverse DNS, and open relay vulnerabilities — all in one place. The free blacklist monitor will alert you if your domain gets listed. If you've ever spent an afternoon chasing "why is my email going to spam," this is the first tool to reach for.

IPInfo

Paste any IP address (or load the homepage to check your own) and get structured data: geolocation down to city/coordinates, ASN and carrier, whether it's a VPN/proxy/Tor exit node, and the organization name. The free API serves 50,000 requests per month — enough to integrate IP context into a personal project or small service. Used internally by Google, Cloudflare, and GitHub for network intelligence.

DNS Leak Test

Runs a series of DNS queries through your current connection and reveals which DNS resolvers actually handled them — exposing whether your VPN is silently leaking your real ISP's DNS servers in the background. A clean result shows only your VPN provider's resolvers; a leak means your traffic metadata is visible to your ISP even though your IP is hidden. Run this every time you set up a new VPN client.

AI Tools

Elicit

An AI research assistant built specifically for academic literature — search 125+ million papers, extract structured data from thousands of PDFs at once, and get a synthesized summary with inline citations from the actual papers rather than hallucinated facts. The "Notebook" feature lets you build a living literature review that updates as new papers match your query. A researcher's alternative to asking a general-purpose LLM that doesn't know which papers exist.

Replicate

Run any of thousands of open-source ML models via a simple API call — image generation (Flux, SDXL), video synthesis, speech-to-text, LLMs, image-to-image, upscalers, music generation — without provisioning a GPU or writing a line of infrastructure. Pay per second of compute, start in five minutes with a cURL command. The fastest path from "I want to run this Hugging Face model" to actually running it in production.

Exa

A web search API purpose-built for AI agents rather than human browsing: returns clean, structured content with sub-180ms response times, uses embedding-based semantic search (not just keyword matching), and strips 90% of the boilerplate from fetched pages before returning them as LLM context. If you're building a RAG pipeline or an agent that needs to look things up, Exa is dramatically cheaper and cleaner to integrate than scraping raw HTML.

Developer Tools

JSON Crack

Paste any JSON, YAML, CSV, or XML and it instantly renders an interactive visual graph of the data structure — nodes you can drag, zoom, and click to explore nested hierarchies. Export as PNG or SVG for documentation. Runs entirely in the browser; nothing leaves your machine. Invaluable for understanding deeply nested API responses at a glance when `console.log` produces 400 lines of brackets.

Mockaroo

Generate up to 1,000 rows of realistic fake test data per download in CSV, JSON, SQL, or Excel — with 180+ data types including names, addresses, emails, UUIDs, product names, IP addresses, credit card numbers (fake), and custom regex patterns. The API lets you automate mock data generation in CI. Free tier covers most needs; much more useful than hand-writing seed files and faster than Faker for quick one-off datasets.

Compiler Explorer

Type source code on the left and watch the compiled assembly appear in real time on the right — switching compilers, architectures, and optimization flags live. Supports 30+ languages (C, C++, Rust, Go, Haskell, Fortran, Swift) and 400+ compiler versions; compare GCC vs Clang output side by side, see exactly what `-O2` does to your loop, or share a permalink to any code/compiler combination. Loved by performance engineers and systems programmers; runs entirely in the cloud so no local toolchain needed.

Windows Software

Rufus

A single-file, 1.5 MB portable executable that creates bootable USB drives from ISO images faster and more reliably than any alternative — supports both legacy BIOS and UEFI, handles Windows, Linux, and DOS, and can write to Bad Blocks while creating the drive. No installer, no ads, no upsell. Downloaded over 400 million times. The one tool you want on a USB stick before the next time you need to reinstall Windows or boot a live Linux distro.

Ventoy

Instead of overwriting your USB stick with a single ISO, Ventoy formats it once and then lets you drop any number of ISO files into a folder — all of them bootable at startup via a menu. Supports 1,100+ tested distros and Windows images; the USB remains usable as normal storage for the rest of the files. Open source with an active community; the only sensible way to maintain a multi-tool rescue USB stick.

CrystalDiskInfo

Reads S.M.A.R.T. health data directly from your hard drives and SSDs and displays it in plain language: current temperature, reallocated sector count, pending sectors, uncorrectable errors, power-on hours, and an overall health status of "Good," "Caution," or "Bad." Free, portable, and lightweight. Run it quarterly and you'll usually get weeks of warning before a drive fails rather than losing everything to a surprise death.

Open Source Software

ntfy

Send push notifications to your phone or desktop from any script, cron job, or server with a single `curl` command — no app SDK, no OAuth, no API keys for the free public server. Subscribe to a topic on the phone app and `curl ntfy.sh/my-topic -d "backup complete"` from anywhere. Self-hostable on your own server for private use. Eliminates the need for Telegram bots or Slack webhooks for personal automation alerting.

Changedetection.io

A self-hosted website monitoring tool that watches any URL for changes and notifies you via 85+ channels (Discord, Telegram, email, ntfy, Slack, webhooks). Handles JavaScript-rendered pages via Playwright/Puppeteer, supports XPath/CSS selectors to watch only specific page elements (a product price, a stock status, a job posting), and stores a diff history so you can see exactly what changed and when. The open-source answer to expensive commercial monitoring services.

Portainer

A clean, web-based GUI for managing Docker containers, Swarm clusters, and Kubernetes without memorizing CLI commands — deploy containers, manage volumes and networks, view logs, exec into shells, set resource limits, and monitor CPU/memory usage from any browser. The Community Edition is fully free and covers everything a homelab or small team needs; Enterprise adds RBAC and SSO. One `docker run` command deploys it; 30 million+ deployments worldwide.

Productivity & Collaboration

Anytype

A local-first, end-to-end encrypted personal knowledge base and workspace — notes, databases, tasks, wikis, and diagrams all stored on your device with optional P2P sync across your own devices. Everything is built from linked "objects" you can connect into a personal knowledge graph, with offline-first access that works on planes and subways. Open source, no vendor lock-in, and the syncing infrastructure never sees decrypted data.

Zotero

The best free reference manager for anyone who reads academic papers: one-click browser imports from 1,000+ databases and journals (pulling PDF, title, authors, DOI, and abstract automatically), annotation and tagging, full-text search across your library, and automatic citation/bibliography generation in 9,000+ styles for Word, LibreOffice, and Google Docs. Stores your entire library in a portable SQLite database with optional free 300 MB cloud sync. Indispensable for students, researchers, and knowledge workers.

Plane

An open-source project management tool designed as a genuine Jira and Linear replacement — issues, cycles (sprints), modules (milestones), pages (wiki), and analytics all in one, with GitHub integration, Slack notifications, and a clean UI that doesn't require a 30-minute onboarding tutorial. Deploy it yourself on any VPS or use the hosted cloud tier. If you've ever been frustrated by Jira's complexity or Linear's price, Plane offers most of the substance at $0 for self-hosters.

Learning & Reference

VisuAlgo

Step-by-step animated visualizations of 24 classic data structure and algorithm topics — sorting (bubble, merge, quick, heap), binary trees, graphs (BFS/DFS/Dijkstra/Bellman-Ford), segment trees, union-find, and more. Developed by NUS faculty and used by CS students worldwide; each module includes lecture slides, interactive mode (step through each operation manually), and auto-graded quizzes. By far the clearest visual explanation of "why does merge sort do that?" available for free.

Missing Semester (MIT)

MIT's free "Missing Semester of Your CS Education" — a course on the practical tools no university class teaches: shell scripting, Vim, Git internals, tmux, SSH, dotfiles, debugging tools, metaprogramming, and security basics. All lectures are recorded and freely available; the exercises are genuinely hard. Designed for students who know algorithms but feel lost when they leave the IDE and open a terminal.

Big-O Cheat Sheet

A single dense reference page of time and space complexity for every common data structure operation (search, insert, delete, access) and sorting algorithm — plus charts visualizing how O(n²) vs O(n log n) actually diverges. No ads, no account, no upsell. The single page you want bookmarked before any technical interview or when you're deciding whether a `dict` or a `list` is the right structure for a 10 million-item dataset.

Fun / Weird Internet

Windows 93

A full fake operating system running in your browser — with a taskbar, draggable windows, a file system, a Paint clone, a working Terminal, a Winamp-style music player, easter eggs, and dozens of tiny apps hidden inside. Endlessly deep and genuinely weird; discovering what's buried in it is half the fun. Built by two French developers as an art project and left running freely for anyone to explore.

This Is Sand

Click and drag to pour colored sand down the screen, watching it pile up and blend in satisfying physics simulation. Switch colors, layer them, and build intricate layered landscapes entirely from virtual sand. No instructions, no goal, no account — just a meditative creative tool that people somehow spend 20 minutes in without realizing. The sharing gallery is full of astonishing artwork made entirely this way.

Every Noise at Once

An algorithmically-generated, continuously-updated scatter plot of every music genre on Spotify — over 6,000 genres placed on a two-axis map (mechanical/organic on one axis, atmospheric/spiky on the other). Click any genre label to hear a 30-second representative sample; click the arrow to see the artists defining it. A single page that makes the entire topology of human music browsable. Built by a Spotify data engineer as a side project; impossible to visit without discovering three genres you didn't know existed. --- *30 sites this week. Found something broken or have a site to suggest? Reply to the email.*