Most developers use blockchains the same way they use AWS. They call the API. They trust the infrastructure. They never think about what's actually running underneath. Running your own node changes that permanently. @inkonchain — 36,800 GitHub stars, MIT licensed, and the official Docker Compose setup for running a full Ink L2 node on Kraken's DeFi chain. Here's what spinning up your own node actually gives you: → Full node — validates and serves the complete chain state without relying on any third-party RPC provider → Archive node — full history of every state since genesis, downloaded via Gelato ChainSnap for fast bootstrap → op-geth op-node — the OP Stack execution and consensus client pair running Ink's canonical chain → Grafana dashboard at localhost:3000 — real-time monitoring, sync status, state root fault detection built in → Healthcheck service — alerts you the moment your node falls out of sync with the reference L2 → .env config — one file for RPC endpoint, L1 beacon, node type, and custom port overrides → git clone docker compose up — node running in minutes → Hardware baseline: 8 CPU cores, 8GB RAM, 100GB disk → MIT licensed — 631 forks, actively maintained Note: op-geth support ends May 31, 2026. Migration to op-reth is the recommended path for production operators. Your node. Your chain. No intermediaries. Discovered on OSSphere : ossphere.dev/inkonchain/node Have you ever run your own blockchain node — and what did it teach you? Drop it below 👇 #Ink #OpenSource #Web3 #Ethereum #BuildInPublic #Layer2 #NodeOperator

The developers who will matter most in 5 years are building their own tools right now. Not waiting for the perfect SaaS product. Not paying $49/month for something they could build in a weekend. Not locked into someone else's roadmap. They're shipping small, focused open source tools that solve their own problems — and discovering that other people have the same problem. The pattern repeats constantly: → A developer builds a CLI tool for their own workflow → Posts it on GitHub with a clean README → Hacker News upvotes it → 10,000 stars in a week → The tool becomes infrastructure This has happened with bat, fd, ripgrep, zoxide, Starship, LosslessCut, ShareX, sniffnet — all of them. None of them started as products. All of them started as one person's personal itch. The open source tools you use daily were built by developers who stopped waiting for someone else to build what they needed. The bar has never been lower to build something real: Rust, Go, TypeScript, Python — pick one. Find the tool you wish existed. Build it. Open source it. ossphere.dev What tool do you wish existed but nobody has built yet? Drop it below 👇 #OpenSource #BuildInPublic #IndieHacker #GitHub #SideProject #OSS #DeveloperTools

A documentation repo with 37,000 GitHub stars. That's not a typo. And it tells you everything about how much the crypto developer community is watching this. @inkonchain is Kraken's dedicated DeFi chain — a Layer 2 built on Optimism's OP Stack, part of the Superchain ecosystem. And the developer documentation is fully open source. Here's what builders are getting access to: → Full smart contract deployment guides — Foundry, Hardhat, and Remix — step by step, production-ready → Node running guide — run your own Ink node from scratch → Transaction fees — how gas works on Ink, explained clearly → Account abstraction — smart accounts and onchain clients → Superchain deployment — SuperchainERC20 bridging and cross-chain → Account abstraction, oracles, VRF, RPC, indexers — all covered → Block explorers, bridges, multisig, security tooling — full stack → InkGPT — AI assistant trained on the docs, built right in → Ink Builder Program — Spark, Forge, Echo programs and office hours → Kraken Verify — identity layer for the chain → Built with Next.js Nextra — PR previews via AWS Amplify → 447 forks — community contributing documentation improvements Built by one of crypto's most trusted exchanges. Secured by Ethereum. Scaled by Optimism. Discovered on OSSphere : ossphere.dev/inkonchain/docs Are you building on any L2 chains in 2026 — and which one? Drop it below 👇 #Ink #Kraken #OpenSource #Web3 #BuildInPublic #Ethereum #Layer2

Five years ago, "I built this with open source tools" was a caveat. Something you said quietly, hoping nobody would ask follow-up questions about reliability, support, or whether it was enterprise-grade. Today it's a flex. Here's what the open source credibility shift actually looks like: → Linear is built on open source infrastructure → Vercel is built on open source infrastructure → Notion uses open source at its core → Figma's real-time engine is built on open source principles → The AI tools reshaping every industry run on PyTorch, Transformers, and CUDA — all open source → The fastest-growing developer tools of 2025 are mostly OSS → YC now asks founders about their OSS strategy, not just their build-vs-buy decisions The conversation has completely flipped. It used to be: "can we trust open source for this?" Now it's: "why would we pay for something the community has already built better?" The developers who understood this shift 3 years ago built faster, shipped more, and spent less. The ones who are figuring it out now are catching up quickly. ossphere.dev What's the most impressive thing you've built primarily with open source tools? Drop it below 👇 #OpenSource #BuildInPublic #Startups #DeveloperTools #OSS #TechStrategy #IndieHacker

Every business has documents. Contracts. Insurance policies. Research papers. Medical records. Financial reports. Legal filings. Unstructured text holding critical information that nobody can query, search, or analyze at scale. Google open sourced the tool that changes that. @GoogleAI — 26,300 GitHub stars, Apache 2.0, and the cleanest approach to structured extraction from unstructured text available today. Here's what a few lines of Python actually give you: → Define what you want to extract using few-shot examples — no complex prompt engineering required → Precise source grounding — every extracted entity mapped back to its exact character offset in the original text → Interactive HTML visualization — highlight exactly what was extracted and where it came from → Reliable structured outputs — schema-enforced via Controlled Generation in supported models → Multi-pass extraction — higher recall on long documents → Parallel processing — handles large documents efficiently → Gemini, OpenAI, and local models via Ollama — your choice → CJK language support — Chinese, Japanese, Korean (v1.1.1) → Community provider plugins — extend to any model backend → MCP server available — plug directly into AI agent workflows → Works across: legal, medical, finance, research, compliance → 1,800 forks, Apache 2.0 licensed Extract contract terms. Pull insurance clauses. Index research findings. Parse medical reports. Query what was unqueryable. Discovered on OSSphere : ossphere.dev/google/langextr… What's the most painful document parsing problem your team has ever faced? Drop it below 👇 #LangExtract #OpenSource #AI #NLP #BuildInPublic #Google #StructuredData

Slack costs $7.25/user/month. For a team of 50 that's $4,350/year. And every message your team sends lives on Slack's servers. In perpetuity. Under their terms. Tesla, Mercedes-Benz, the US military, and hundreds of universities chose a different path. @iantien 36,000 GitHub stars, 8,500 forks, and the open source collaboration platform built for teams where data sovereignty is non-negotiable. Here's what running your own Slack actually gives you: → Channels, threads, DMs, file sharing — everything Slack has → Playbooks — repeatable checklists, automations, and retrospectives for your most critical workflows → AI Agents plugin — multiple LLMs directly in your workspace → GitHub, GitLab, Jira, Bitbucket integrations — devops-native → Jitsi and Zoom call integration — video without leaving chat → Mobile: iOS and Android, Desktop: Windows, macOS, Linux → 20 language translations — community-maintained globally → Docker deployment: running in under 5 minutes → E0 tier: free forever for self-hosted, unlimited users → Enterprise tiers for SSO, compliance, and advanced controls → Apache 2.0 licensed — your data, your server, your rules Europe leads global Mattermost adoption — over 60% of tracked deployments. The pattern is clear: where data sovereignty matters, Mattermost is the answer. Discovered on OSSphere : ossphere.dev/mattermost/matt… Is your team's communication data sitting on someone else's server right now? Drop it below 👇 #Mattermost #OpenSource #SelfHosted #Slack #BuildInPublic #DataSovereignty #TeamCollaboration

The open source community has a word for developers who take but never give back. They don't say it out loud. But they notice. Here's the uncomfortable truth about OSS consumption: The average developer uses dozens of open source packages daily. Most have never filed a single issue. Never opened a PR. Never left a comment. Never starred a repo they actually use. That's not a crime. But it is a pattern worth examining. Because open source only works if the ratio of takers to givers doesn't become completely unsustainable. Here's what "giving back" actually looks like at every level — no senior engineering title required: → Star the repo — takes 2 seconds, signals real usage → File a clear, reproducible bug report — worth more than you think → Answer someone else's issue — you know this; share it → Improve one sentence of documentation — it's always there → Share the project with your network — distribution is valuable → Sponsor a maintainer — $5/month is real money to a solo dev → Write about how you use it — tutorials compound forever → Open a PR — even a typo fix changes your relationship with the code The open source tools you depend on were built by people who gave something back. ossphere.dev What was the last OSS project you contributed to in any form? Drop it below 👇 #OpenSource #ContributeToOSS #GitHub #BuildInPublic #DeveloperCommunity #OSS #GiveBack

Google Analytics 4 is free. It's also confusing, cookie-hungry, forces a consent banner on every visitor, and sends your users' data to Google. 37,000 developers decided there's a better way. @caozilla — 37,000 GitHub stars, 7,200 forks, MIT licensed, and shipping faster than ever in 2026. Here's what privacy-first analytics actually looks like: → No cookies — zero consent banner required, GDPR compliant by design → Real-time dashboard — live visitor counts and pageviews instantly → Custom event tracking — button clicks, form submissions, conversions → Session Replay — watch real user sessions replayed in the browser → Heatmaps — see exactly where visitors click and scroll (May 2026) → Boards — build custom dashboards on a flexible drag-and-drop canvas → Web Vitals tracking — Core Web Vitals performance metrics built in → Multi-value filtering — slice your data any way you need → White-labeling — your brand, your analytics, your clients → Share pages — send a read-only dashboard link to your team → Self-host: docker compose up -d — running in 2 minutes → Umami Cloud — managed hosting if you'd rather not self-host → Full REST API — programmatic access to all your analytics data Your users' data belongs to you. Not to Google's ad targeting machine. Discovered on OSSphere : ossphere.dev/umami-software/… Are you still sending your users' analytics data to Google — or have you made the switch? Drop it below 👇 #Umami #OpenSource #Analytics #Privacy #BuildInPublic #GDPR #GoogleAnalytics

The most valuable skill in open source isn't coding. It's knowing how to ask for help correctly. Most developers open an issue like this: "It doesn't work. Please fix." The maintainer reads it, sighs, and moves to the next one. Here's what a great issue actually looks like: → One sentence: what you expected to happen → One sentence: what actually happened → The exact version of the library you are using → A minimal reproduction — the smallest possible code that triggers the problem, nothing else → Your environment: OS, runtime version, relevant config → What you already tried — so the maintainer doesn't repeat it → No demands. No "this is blocking me urgently." They know. They always know. A well-written issue does something powerful. It makes the maintainer's job easier. And the maintainer's job is already hard. The developers who file great issues get responses faster, get fixes shipped faster, and often get invited to contribute. Not because they're special. Because they respected someone's time. ossphere.dev What's the worst issue format you've ever seen filed? Drop it below 👇 #OpenSource #GitHub #ContributeToOSS #BuildInPublic #DeveloperCommunity #OSS #SoftwareEngineering

Every time you use a modern writing tool on the web — a Notion-like editor, a documentation platform, an AI writing assistant — there's a high chance this is underneath it. @PhilipIsik built the headless rich text editor that powers the web's best writing experiences. tiptap — 37,200 GitHub stars. v3.25.0 shipped last week. Built on ProseMirror. Designed for web artisans. Here's what "headless" actually means for your product: → Zero UI included — your design, your brand, zero class overrides → Framework-agnostic — React, Vue, Svelte, vanilla JS, all supported → 100 extensions — bold, tables, drag handles, code blocks, mentions, slash commands, file uploads and far more → Extension architecture — every feature is modular and swappable → Build custom nodes and marks — extend anything in the schema → Collaborative editing via Hocuspocus — Yjs CRDT backend, real-time multiplayer, open source and self-hostable → AI Agent extension — drop AI assistance directly into the editor → Content AI — generation, completion, transformation built in → Comments, versioning, document conversion — Pro extensions → tiptap-ui-components — ready-made React templates for fast starts → PHP package — server-side content processing without a headache → MIT licensed core — 3,000 forks, shipping weekly The editor your users will love. The codebase your team can actually own. Discovered on OSSphere : ossphere.dev/ueberdosis/tipt… What's the hardest part of building a great rich text editor experience? Drop it below 👇 #Tiptap #OpenSource #RichTextEditor #React #BuildInPublic #ProseMirror #WebDevelopment

December 2022. ChatGPT had just launched. A community of thousands asked: what if we built this ourselves? Open source. Community-trained. No corporation in control. @ykilcher coordinated the effort. 37,500 GitHub stars. 161,000 messages. 35 languages. Hundreds of contributors. Zero VC funding. The project didn't finish. What it left behind changed AI research forever. Here's what Open-Assistant actually accomplished: → oasst1 dataset — 161,000 human-written messages in 35 languages, the most important open RLHF dataset ever crowdsourced → Proved that community-driven AI training is possible at scale → Demonstrated the full RLHF pipeline — from data collection to reward modeling to fine-tuning — in a fully open repo → Built a data collection platform used by tens of thousands of volunteer annotators worldwide → Created infrastructure for rating, ranking, and labeling AI responses that entire research community has since studied → The oasst1 dataset is still actively used to fine-tune LLMs across dozens of open source projects today → Apache 2.0 licensed — 3,300 forks, studied worldwide The model didn't beat GPT-4. The mission didn't complete. But the dataset trained models that are still running in production systems across the world right now. That's not failure. That's how open source moves science forward. Discovered on OSSphere : ossphere.dev/LAION-AI/Open-A… Which abandoned OSS project do you think left the biggest legacy behind? Drop it below 👇 #OpenAssistant #OpenSource #LLM #RLHF #BuildInPublic #AI #LAION

Nobody talks about the moment open source actually clicks. You've been using a library for months. You hit a bug. The docs don't cover it. Stack Overflow has nothing. So you open the source code. And for the first time — you read it. Not to contribute. Not to learn. Just to find the answer. And you find it. In 20 minutes. Inside 3 files. That's the moment. The moment you realize the source code is the documentation. That the answer was always there — you just never thought to look. After that moment, everything changes: → You stop being afraid of "how does this actually work" → You start trusting your own ability to find answers → You file better bug reports — with line numbers, not vibes → You write better code — because you've seen good code up close → You become a more confident contributor — because now you've been inside the thing Every developer has this moment. Most haven't had it yet — not because they're not ready, but because nobody told them they were allowed to just open the file and look. You're allowed. The source code is right there. ossphere.dev When did open source first click for you? Drop it below 👇 #OpenSource #BuildInPublic #SoftwareCraft #GitHub #DeveloperGrowth #CodeReading #OSS

Cursor is beautiful. GitHub Copilot is convenient. But if you live in the terminal and want every AI edit to be a clean, reversible Git commit — there's only one tool. The most popular terminal-based AI coding agent in existence. Here's what makes it different from everything else: → Git-native by design — every AI edit auto-commits with a meaningful conventional commit message. Always. → /undo undoes the last AI change instantly — one command, no fear, no manual git revert → 100 LLM providers via LiteLLM — Claude, GPT, Gemini, DeepSeek, Mistral, local models via Ollama — you choose → Architect mode — expensive model plans the feature, cheap model writes the code — cuts API costs significantly → Tree-sitter repository maps — understands your entire codebase structure, not just the open file → Handles monorepos with hundreds of thousands of lines — tested → Auto-runs lint and tests — fixes failures without being asked → Watch mode — bridges terminal and Vim, Emacs, or Neovim → Works anywhere: local machine, remote server, Docker container → Community convention files — project-specific AI behavior → Apache 2.0 licensed — 4,500 forks, free forever No GUI. No extension. No subscription. Just your terminal, your LLM, and clean Git history. Discovered on OSSphere : ossphere.dev/Aider-AI/aider Terminal or IDE — how do you prefer to work with AI coding tools? Drop it below 👇 #Aider #AICode #OpenSource #Git #BuildInPublic #Terminal #DeveloperTools

The best software interview question I ever heard wasn't about algorithms. It was: "Tell me about an open source project you use daily and explain one thing about how it works internally." Most candidates couldn't answer it. Not because they weren't smart — but because they had never looked inside the tools they depend on every single day. Here's what looking inside actually teaches you: → Why abstractions exist — not just what they do → How tradeoffs get made under real constraints → What "good enough" looks like in production code → How experienced engineers name and organize things → Why your ORM generates that SQL — and when to bypass it → How your HTTP client handles retries, timeouts, redirects → What your test runner does between "run" and "pass" → Why your build tool makes the choices it makes The developers who understand their tools at one level deeper than the documentation write better code with those tools. They also debug faster. Contribute better PRs. And give more interesting interview answers. GitHub has the source code for everything you use. The README is the surface. The src folder is the education. ossphere.dev What tool do you use daily but have never looked inside? Drop it below 👇 #OpenSource #SoftwareCraft #GitHub #BuildInPublic #DeveloperGrowth #CodeReading #SoftwareEngineering

In 2015, a small team asked one question: What if deploying to AWS Lambda was as simple as one YAML file and one command? 46,900 GitHub stars later — Serverless Framework is still the answer most teams reach for first. @goserverless has been shipping since before "serverless" was even a common word. And in January 2026, it got significantly more powerful. Here's what one serverless.yml gives you: → Deploy to AWS Lambda, Azure Functions, Google Cloud Functions, and more from a single configuration file → Auto-scale to any load — zero management, zero capacity planning → Zero cost when idle — pay only for what actually runs → Managed Instances — new in Jan 2026, persistent serverless state → Durable Functions — long-running workflows, built in → Built-in AppSync and Prune plugins — shipped January 2026 → AWS Login and SSO support — one command auth setup → Local invocation and testing — debug before you deploy → Stages and environments — dev, staging, prod from one config → 11,500 example templates across every language and use case → Hundreds of community plugins — extend anything → Free for individuals and orgs under $2M revenue → MIT licensed core — 5,700 forks, updated yesterday serverless deploy. That's the whole workflow. Discovered on OSSphere : ossphere.dev/serverless/serv… Are you running serverless in production — and would you go back to managing servers? Drop it below 👇 #Serverless #OpenSource #AWS #Lambda #BuildInPublic #CloudComputing #DevOps

GlassWire charges $29–$99/year to show you what's happening on your own network. Wireshark is free — but built for network engineers with patience and a manual. @GyulyVGC built the third option. sniffnet — free, beautiful, built for everyone. 38,100 GitHub stars. 401,000 downloads. Built in Rust by one developer. Here's what it shows you in seconds: → Real-time traffic rate chart — incoming and outgoing, last 30 seconds, live → Network hosts dashboard — hostname resolution, not just raw IPs → Service and protocol breakdown — see exactly what's talking and where it's going → Filter by adapter, protocol, IP, and port — drill into anything → IP blacklist monitoring — get alerted when blacklisted IPs connect → Custom notifications — data threshold exceeded, new favorites activity, blacklisted IP detected — with alert sounds → Remote notifications via webhook — integrate with any system → PCAP export and import — full capture reports, standard format → Pause and resume packet captures — without losing your session → Custom themes — TOML-based, plus Dracula, Gruvbox, Nord, Solarized → Windows (x64 ARM64), macOS, Linux (AppImage, DEB, RPM) → GPL-3.0 licensed — 1,600 forks, v1.4.2 shipped April 2026 You deserve to know what your computer is sending. Sniffnet makes that knowledge comfortable. Discovered on OSSphere : ossphere.dev/GyulyVGC/sniffn… Do you actively monitor your network traffic — or do you just hope for the best? Drop it below 👇 #Sniffnet #OpenSource #NetworkMonitoring #Privacy #BuildInPublic #Rust #CyberSecurity

The open source tools developers actually use daily rarely trend on GitHub. They just quietly work. Every single day. For years. Here's the unglamorous list — the tools in every serious developer's setup that nobody tweets about: jq — parse and transform JSON from the terminal. Every backend developer uses it. Almost nobody talks about it. fzf — fuzzy finder for the terminal. Once you install it, you forget how you lived without it. tmux — terminal multiplexer. Still running the same session you started three days ago. htop — interactive process viewer. The first thing installed on every new server. rsync — file synchronization that just works. The reliability benchmark everything else is measured against. httpie — curl but readable. The first tool that made HTTP requests feel human. ncdu — disk usage analyzer for the terminal. Found 40GB of lost Docker layers in 30 seconds. entr — run commands when files change. The simplest watch tool ever built. mkcert — local HTTPS in one command. Solved the problem every developer hit every week. These tools don't have launch tweets. They have 15-year changelogs and zero marketing budget. ossphere.dev What terminal tool could you absolutely not work without? Drop it below 👇 #OpenSource #CLI #Terminal #DeveloperTools #BuildInPublic #Linux #SoftwareEngineering

Snagit: $62.99 one-time. Camtasia: $299.99. Loom: per-seat pricing that scales awkwardly. Droplr: subscription. ShareX: $0. Forever. No ads. No watermarks. No paid tiers. For 18 years straight. ShareX/ShareX — 38,000 GitHub stars, maintained since 2007, and still the most powerful free screen capture tool on Windows. Here's what one hotkey actually gives you: → Fullscreen, window, region, and scrolling capture — all modes → Webpage capture — full-length screenshots of any site → Screen recording — GIF and video, directly from the capture workflow → OCR — extract text from any screenshot in seconds → Image editor — annotate with shapes, blur, text, arrows, highlights → 80 upload destinations — Imgur, Google Drive, Dropbox, FTP, custom servers, and your own self-hosted backend → URL shortening built into the upload workflow → After-capture tasks — watermark, QR code, clipboard, browser open → Fully customizable hotkeys — one key, your entire capture workflow → Color picker, ruler, hash checker, DNS changer — all included → .NET 9.0, C#, GPL-3.0 — 3,800 forks, updated June 2026 Microsoft's best Utility App 2022. The Guardian's top Snipping Tool replacement. Lifehacker's best screenshot tool for Windows. Two developers. 18 years. Still shipping. Discovered on OSSphere : ossphere.dev/ShareX/ShareX What's the screen capture tool you use daily — and are you actually happy with it? Drop it below 👇 #ShareX #OpenSource #Windows #Productivity #BuildInPublic #ScreenCapture #DeveloperTools

There's a conversation happening in every engineering team right now. "Should we open source this?" It's a harder question than it looks. Here's the honest framework: Open source it if: → The problem is universal — other teams are solving it too → Distribution matters more than defensibility for this component → External contributors will find bugs you never would internally → You want to attract engineers who care about the problem → The component isn't your core competitive advantage → You can commit to maintaining it publicly — indefinitely Don't open source it if: → It's genuinely proprietary — the code IS the moat → You can't resource the maintenance burden it creates → The codebase has security or compliance landmines → You're doing it for marketing reasons with no real commitment → You'll abandon it in 18 months when priorities shift The graveyard of "open sourced and abandoned" repos has done more damage to developer trust than closed source ever did. Open sourcing something and then going quiet is worse than never open sourcing it at all. The best decision isn't always "open source everything." It's "open source the things we'll actually maintain." ossphere.dev Has your team ever open sourced something? What happened? Drop it below 👇 #OpenSource #EngineeringCulture #BuildInPublic #SoftwareStrategy #OSS #TechLeadership #GitHub