New Folo is out: Social Media posts are marked read as you scroll again, changes made offline sync once you're back, and iPad screens fit every window.

New Folo is out: picture feeds on Android stay light and fast, Discover finds the feeds behind any website, and forms no longer drop your edits, plus hundreds of fixes.

New Folo is out: your settings follow you to every device again, long reads get text-to-speech that holds up, and every screen now speaks Chinese, Japanese and French.

Folo now syncs only what changed, so your feeds stay current with a fraction of the requests 🔄

🧡

💪 New Al models

🧡 Folo now lets you catch up on unread entries oldest first, keeps your place when switching timelines on mobile, and brings .deb packages to Linux.

The best updates aren’t always the loudest. This release makes Folo safer, clearer, and more polished across desktop and mobile—from integrations and error messages to dark mode and visual details. ✨

Which fix or upgrade are you most excited about in the latest Folo release? 👀

Folo Discover today is basically a map of what builders are following: Anthropic Research, OpenAI News, GitHub Trending, Hacker News, Karpathy, Stripe Engineering, Schneier on Security, NASA APOD… AI, open source, engineering, security, tech news, and curiosity, all in one feed. Don’t wait for the algorithm to decide what matters. Follow the sources that do. Discover your next high-signal feeds on Folo.

A small but very practical Folo update. This one makes everyday reading feel smoother: 📖 fewer accidental mark-read moments 🧩 cleaner code blocks 🎧 better desktop reading and TTS behavior Desktop v1.10.0 and mobile v0.5.5 are rolling out now.

Need to optimize my World Cup information setup. What are the best underground or tier-1 sources for data/news that desperately need an 🟧RSSHub feed? Drop your favorites, trying to build a master list ⚽️

Wake up. Open 🟠Folo. Check what’s happening in the world before checking what’s happening on X. Reuters. BBC. NYT. Foreign Affairs. All in one place.👀

2024: “RSS is dead.” 2026: The internet is scattered across apps, channels, repos, forums, newsletters, and group chats. Keeping up has become a full-time job. But we’re still here 👀 🟦 RSSHub: turn scattered information into open, accessible feeds. 🟠 Folo: turn those feeds into something worth following.

One thing we’ve realized while actually building products: Users rarely look at your model and say “the reasoning here is sub-optimal.” Instead, they get pissed off when the product feels like a black box. It’s always the same annoying stuff: hitting quota limits out of nowhere, sudden latency spikes, or the system silently dying without any error logs. As AI products mature, you quickly realize that users value reliability way more than a slightly higher benchmark score. Intelligence is cool and all, but in production, “boring and predictable” is the ultimate feature.🟦

Building a flagship product for mass adoption is usually less glamorous than one big release. It’s the small improvements: better speed, smoother UX, more reliable everyday use. Folo keeps getting better, version by version. 🟦🟠

In 2010, Laszlo Hanyecz just wanted some pizza. In 2026, AI devs just want High-Fidelity inputs that don’t catastrophically break their RAG workflow. Who’s crying more? 🥲 If your upstream data pipeline has zero telemetry, you’re still paying 10,000 BTC in technical debt. Time to upgrade to 🟦RSS3’s machine-readable information. Workers are cooking the data. You enjoy the pizza. 🍕⚡️

Scaling AI infra is no longer just a model capability problem, but increasingly a data integrity problem. Today, many AI systems still rely on fragmented schemas, opaque APIs, and inconsistent upstream sources. A small change in formatting or structure can quietly break downstream reasoning. 🟦RSS3 focuses on making open information more structured and machine-readable. Through open Workers and standardized processing pipelines, raw signals across protocols and platforms can be transformed into consistent, reusable metadata that applications and AI systems can actually work with. Reliable AI systems start with better inputs, rather than bigger models.

Ethereum migration is now complete 💙 VSL users can access your migrated assets here 🟦 → Link rss3.io/migration