v1.39 is live 🎉 The biggest one yet: home, games and the plaza are now a single map - Giggle World. Walk around, run into other learners, react with emotes, invite a friend to join the battle. Here's everything that shipped 🧵 1/ Giggle World 🗺️ No more separate tabs. Your home, the game hall and the plaza are on one map now. The plaza is properly multiplayer - you'll see other learners on screen, pick up emotes and props, tap someone's profile card to visit their place or start a PK battle. 2/ Phone Calls with Max 2.0 ☎️ Calls used to follow a fixed script. Now they're built from live scenarios with real-time speech so when a kid says something unexpected, Max actually responds to it. It's a conversation, not a recording. 3/ Meet the parrot Kiki 🦜 Say a word, Kiki says it right back. Get it right and you collect the little creature that goes with it. Repetition is how pronunciation sticks - this just makes the repetition something kids want to do. 4/ Bubble Words 🫧 New solo game in the game hall. Hear a word, pop the right bubble. Three difficulty levels, combo streaks during bubble waves, and national + global leaderboards. Fair warning: your kid will beat your score. 5/Plus🛠️ • Story characters you can talk to in English, as NPCs • A new audio-matching question type in unit reviews • Drawing now works fully offline • Faster loading on weak connections Free forever, for every child. Update and go say hi in the plaza 💛

v1.0.29 is here! 🐀 For radio enthusiasts, this is probably the most important update in a while, as it brings stability across the board for BLE connected radio devices. For the full changelog, check it out on GitHub:

Hi3D V1.0 (formerly Sparc3D) is building on last year's 3D research with roughly 2 million training assets, up from 500,000, according to the team. Hi3D (@Hi3D_AI) has now released the code and weights of V1.0, a model the team says has already been validated in real production workflows. In the team's technical report, V1.0 leads the compared models on multi-view fidelity and detail richness, with a 90% slicing success rate. For anyone building 3D assets, the promise here is preserving the small parts and unseen sides that a front-view render can hide. Release (try it free online now, no deployment needed): GitHub:

EverOS v1.5.1 is here. A small open-source update for a frustrating problem: memory searches queuing up under load. We increased the default query-embedding concurrency from 4 to 16, giving hybrid and vector search more room to handle concurrent requests. In our five-minute test with 10,020 indexed entries at 10 requests per second: • p95 latency dropped from 101s to 1.45s • All 3,000 requests were served No migration needed, just upgrade. If your embedding endpoint has tight rate limits, tune the concurrency to match. Prefer to build without managing the infrastructure? EverOS Cloud brings managed memory to your existing agent workflows. Save useful context after a task, retrieve it before the next, and give your agents access to user preferences, business knowledge, and past experience through APIs. Start with one recurring workflow and let useful context accumulate across runs. Subscribe to EverOS Cloud Pro now and get 3× the standard monthly usage for your first 3 months. Cloud:

Bagel V1.5.0 is here! 🥯 A couple of updates to make your Bagel experience even better 👇 → Get notified when people you follow place Combo orders → Getting ready for an upcoming campaign with @Polymarket Update your Bagel app and stay tuned. More fun is coming!