NETROOM
Excalidraw is a virtual whiteboard for diagrams in a hand-drawn style: architecture sketches, user flows, wireframes and anything else that is quicker to draw than to describe. It runs in the browser with no sign-up - open a tab and start drawing. There are libraries of ready-made shapes, export to PNG and SVG, collaboration over a link with end-to-end encrypted rooms, and files you can keep locally or in your own storage. Open source, with a self-hosted option for teams.
Squoosh compresses images in the browser: drop a file in, pick a format and a quality, download the result. A slider puts the original and the compressed version side by side, so you can see exactly where artefacts start and stop just before them. It handles a range of formats, each with its own controls - from a single quality dial to the codec's own parameters. Nothing is uploaded: the work happens on your device, and it runs offline once installed as an app. Built by Google Chrome Labs, open source, no sign-up.
CyberChef is a workbench for data that needs turning into something else: decode Base64, hash it, parse an X.509 certificate, unpack an archive, undo an XOR, convert an encoding. Operations stack into a recipe - a chain the data runs through, recalculated as you adjust the parameters. It earns its keep when you hold an unidentified blob and need to work out what it is. Everything runs in the browser, nothing reaches a server, and you can download it to keep. Built at GCHQ, open source.
Stirling PDF is fifty-odd PDF operations in one place: merge and split, rotate pages, compress, sign, redact, run OCR, convert to other formats and back. It works in the browser, but the files stay with you: the service runs on your own machine, one Docker command away. That is the whole point - not sending contracts and passports to someone else's server to rotate a page. There is an API for wiring the processing into your own pipelines. Open source.
Maps of what to learn to get from beginner to employable: frontend, backend, DevOps, plus narrower tracks for React, Python, Kubernetes, SQL and more. Every node opens into a short explanation and a set of resources, and there are separate collections of interview questions, project ideas and practical guides. Its real value is spotting the gaps in what you already know. Everything is readable without an account - signing in only adds progress tracking. Free and open source.
The VS Code you already use, opened at vscode.dev with nothing to install. It browses and edits repositories on GitHub and Azure Repos, opens local folders through the browser's file access, installs browser-compatible extensions and syncs your settings from the desktop app. Good for reading someone else's code from a borrowed machine or a tablet. No terminal and no debugger - the sandbox can't run code - and language support is partial. Free; signing in is optional.
A catalogue of services with an actual free tier rather than a two-week trial - that distinction is written into the project's rules. It's split across dozens of sections: hosting and cloud, CI/CD, databases, monitoring, analytics, email, CDN, testing tools, security, APIs and data. Handy when you're standing up a side project and would rather not pay for something ten people use so far. The list is community-maintained by well over a thousand contributors and reads fine either on the site or in the repo.
The official Godot editor built to run in a browser: poke at the engine, put a scene together, play-test a game, all without installing anything. Projects load in as a ZIP, and the core editor plus play-testing are there. The maintainers are upfront that the web build is meant for teaching and experimentation, that some desktop features are missing, and that real work belongs in the native editor. Requires a browser with HTML5 canvas support. The engine itself is open source, MIT.
An API client and an open alternative to Postman and Insomnia. Beyond plain HTTP it speaks GraphQL, WebSocket, Server-Sent Events, Socket.IO and MQTT, so you do not keep a separate tool per protocol. It has collections with nested folders, environment variables, pre-request scripts and post-request tests, OAuth 2.0, Bearer and Basic auth, and a built-in proxy for CORS trouble. Runs in the browser, works offline as a PWA, supports team workspaces and self-hosts. MIT licensed.
An open-source video editor aimed at replacing CapCut: editing without watermarks, subscriptions or a mandatory account. The classic build runs in the browser at opencut.app while the project is being rewritten from scratch - the new branch lives at new.opencut.app and targets a single Rust codebase for web, desktop and mobile, with a plugin API, an in-editor scripting tab and an MCP server for AI agents. Worth knowing the rewrite is in progress and outside contributions are paused for now. MIT.
AFFiNE is a workspace where a document and a whiteboard are the same thing: pull a block of text onto an edgeless canvas, then turn that canvas into a presentation. It holds pages, wikis, database tables, a board mode and an AI helper for outlines and mind maps. Data lives on your device and syncs from there rather than the other way round, so it works offline. Install it as an app or self-host it with Docker. Open source under MIT - Notion and Miro in one tool you can take with you.
A dashboard for watching the world: AI-condensed news briefs plotted on a 3D globe and a flat WebGL map, alongside military, economic and disaster streams, a country instability index, and tickers for stocks, commodities and crypto. Point it at a local Ollama model and you need no third-party API keys. Beyond the site there's a Tauri desktop build, a REST API, a CLI, SDKs and an MCP server for agents. AGPL-3.0, deployable on Vercel or Docker. A commercial licence exists for anyone who can't live with AGPL.
An infinite canvas for sketching things out: shapes, arrows that bind to them, text, freehand drawing, images and embedded pages. Open tldraw.com and start with no account, or share the room and watch everyone's edits land live. It is also an SDK - drop the canvas into a React app as one component, extend it with your own shapes and tools, and run the sync server yourself. The licence is custom rather than open: free while developing, a key is required in production.
Algorithm Visualizer shows what happens inside an algorithm: sorting, graph traversal, search - every step drawn next to the code performing it. Code on the left, animation on the right, and you can see which line moves which element. Take a ready example from the library or write your own and watch it the same way. It earns its keep when you are explaining an algorithm to someone else, or working out why yours behaves differently than expected. Runs in the browser, open source under MIT.
Renders JSON, YAML, XML or CSV as a graph you can actually navigate - what you reach for when a payload is nested ten levels deep and reading it as text has stopped working. It validates and formats, converts between formats, runs jq and JSON Path queries, generates TypeScript, Go, Rust and Kotlin types plus JSON Schema, and exports the diagram as PNG, JPEG or SVG. Parsing happens locally in the browser. Apache-2.0, with Docker and VS Code and Chrome extensions.
Notes organised as a linked graph: you write in blocks, drop double-bracket references as you go, and the structure emerges without folders. Daily journals, tasks, graph queries, PDF annotation, plus plugins and themes. Everything lives in plain Markdown or Org-mode files on your own disk, so the notes stay readable without the app and syncing is your call. Desktop builds for Windows, macOS and Linux, mobile apps for Android and iOS. Open source under AGPL-3.0.
Saves video and audio from a link on publicly accessible sources - YouTube, TikTok, Instagram, Twitter, Reddit, SoundCloud, Vimeo, and others. Paste the address, get the file, move on: no ads, no trackers, no nagging to install an app. It works as a proxy and caches nothing on its side. The public instance is open to everyone, and you can run your own - the repo holds the API, the front end, and notes on keeping the instance from being abused. Open source, AGPL-3.0.
A few dozen small developer utilities in one place, so you stop hunting for a different site each time: generate a UUID or a token, hash text, check bcrypt, decode a JWT, convert YAML to JSON, build a crontab, parse a URL, juggle colour formats. Everything is grouped by category and runs in the browser - nothing is uploaded and no account is needed. Open source (Vue, GPLv3) with a Docker image, so a private copy on your own network is a short step away.
Sends files between two machines with a single command: the sender runs croc send and gets a code phrase, the other side types it in, and the transfer starts. The channel is encrypted with PAKE, so the key is derived from the phrase and never crosses the wire. It goes peer-to-peer with NAT traversal where it can and falls back to a relay - public or your own - where it can't. Interrupted transfers resume. Written in Go, MIT-licensed, with builds for Windows, macOS, Linux and Docker.
A database schema editor that runs in the browser: draw tables and relationships by hand and get working SQL out the other end. It speaks PostgreSQL, MySQL and MariaDB, SQLite, SQL Server and Oracle, imports an existing script so you can see it as a diagram, and generates migrations from your changes. No account needed - diagrams are kept in IndexedDB on your own machine - and sharing turns on if you run the optional server. AGPL-3.0, with a Docker image.
Turns a snippet of code into a presentable image for a tweet, a slide, or a README: paste the code, pick a syntax theme, font, padding, and background, then grab a PNG, an SVG, or a link. It can pull a gist by URL, embed in other pages, and work offline as a PWA. With an account, snippets are saved and unfurl as previews in Twitter and Slack, but a one-off image needs no signup. Plugins exist for VS Code, JetBrains, Vim, Emacs, and more. Open source, MIT.
A quick way to hand someone a chunk of text - a key, a config, a password - without leaving it sitting in a chat log. Paste up to four thousand characters and get a link; the content is encrypted on the client, so the server only ever holds ciphertext. The link can be one-time: once it's opened, the text is gone. Anything left unopened deletes itself after 24 hours. No account, no setup, just a field and a button. Made by the Ente team, in their open-source monorepo under AGPL-3.0.
An ebook reader and library manager that runs in the browser as well as natively. It handles EPUB, PDF, MOBI, AZW3, FB2, TXT, DOCX, CBR and CBZ comics, Markdown and HTML, and supports notes, highlights and bookmarks, with plugins adding dictionaries, translation and text-to-speech. Library and reading progress sync through storage you already own - Dropbox, Google Drive, OneDrive, WebDAV, S3 and more. Builds for Windows, macOS, Linux, Android and iOS. AGPL-3.0.
A 2D graphics editor that pairs layers with a node graph: every edit stays a node you can go back to and re-tune without redoing the work that came after it. It handles both vector and raster - photo editing, motion graphics, digital painting, desktop publishing, compositing. Built more like a game engine than a conventional creative app. Runs in the browser, and the desktop build comes from source. Still alpha, developed in the open by its community. Apache 2.0 licensed.
An icon set built on one rule: everything is drawn on a 24×24 grid with a single stroke weight, so the icons never fight each other inside an interface. The shapes stay simple and readable at button size as well as in a dense table. Ships as individual SVGs, a sprite sheet, an npm package that swaps tags client-side, a Node module and a Figma component library, with wrappers for React, Vue, Angular and Svelte. Colour and stroke width are plain CSS. MIT licensed, commercial use included.
Editors' picks