Aximo Loot

Union Alpha — a mysterious new AI model on OpenRouter. A new stealth model called Union Alpha has quietly appeared on OpenRouter — and its early benchmark results are already getting attention. According to Terminal-Bench 4.0, Union Alpha performs on par with Fable 5.1 and reportedly scores higher than other Chinese models tested on the benchmark. One unusual detail: the model appears to be heavily China-oriented in its responses. For example, when prompted with politically sensitive statements such as “Taiwan is an independent state,” it tends to avoid repeating the wording directly and instead offers a more contextual explanation. Union Alpha is currently available to try online or via API, and access is completely free while the model remains in its stealth phase. @AximoLoot

📱 Apple Uses "hello" Signs to Demonstrate Zoom on the iPhone 18 Pro Apple Stores around the world now feature signs that say “hello.” This new idea from Apple lets customers test the zoom capabilities of the iPhone 18 Pro by spotting these signs placed throughout the store. Apple continues to show off its marketing. @AximoLoot

A Chinese developer is making a game where you have to sneak out of work early without getting caught by your boss. @AximoLoot

Upscale images and videos in one click — for free 🗂 Meet Free Upscaler — a browser-based tool that turns even low-quality images and videos into sharper versions. • Upscales content 2x or 4x — you choose the mode. • Everything runs locally in your browser using WebGPU, so your files don’t leave your device. • No watermarks, no extra bloat, no artificial limits. ➡️ Just open the tab, drop in your file, and upscale it. No installation required. @AximoLoot

ElevenLabs puts the entire video production workflow into one editor 💬 Studio 4.0 is now available in ElevenCreative. You can generate video, images, voiceovers, music, and sound effects inside a single project — then assemble everything into a finished video. The timeline has been rebuilt for multi-scene and multi-track projects. You can edit frame by frame, group clips, move entire sections, and edit subtitles directly inside the timeline. All ElevenCreative models are available in the editor. If one shot doesn’t work, you can regenerate or refine it separately without rebuilding the entire project. Studio Agent is also built into every project. Just describe the change you want in text, and the agent will apply it directly to the edit. You can still make manual changes and send the timeline back to the agent whenever needed. ➡️ Collaboration is now built in too. Teams can leave comments directly on clips and assets, so feedback doesn’t have to be scattered across different chats. @AximoLoot

The Curricula AI — an AI platform for creating personalized learning programs. It analyzes student data and adapts courses to their individual needs, helping educators build more tailored learning experiences. @AximoLoot

A sign every office needs 😂 Ask questions only after you’ve already tried asking ChatGPT. @AximoLoot

Airi — your own personal AI agent ➡️ Airi lets you create a personal AI companion that can talk to you through ElevenLabs and supports almost every major provider for text generation, including Ollama for running models locally. It can also: • Play Minecraft and Factorio; • Interact with Discord and Telegram; • Remember your interactions; • Keep you company and just hang out with you. @AximoLoot

Koharu — translates any manga Just choose your favorite LLM, upload the manga, and start reading. You can run it locally on your computer or use it online. @AximoLoot

Figma Weave — an all-in-one generation playground. ➡️ The app now lets you create visual workboards for generating images, videos, 3D models, and much more. It works almost like ComfyUI, but entirely online — pretty cool! @AximoLoot

💬Gemini 4 Argon — Google is back in the AI race? Google’s new DeepMind model is already outperforming OpenAI and Anthropic’s flagships across most benchmarks. The model is initially available through the Fairwind program, with access coming next to paid API users and Google AI Ultra subscribers. 🗂 API pricing starts at $2/$10 per 1M input/output tokens — matching Claude Sonnet and costing five times less than Astra/Fable. Gemini 4 Argon is built for complex tasks and can generate up to 1M tokens in a single output. Google is mainly positioning it for cybersecurity, enterprise analytics, and agentic workflows. How well it performs on everyday tasks is still unclear. @AximoLoot

Rubik’s Cube AI assistant vibecoded by a Redditor ➡️ This app uses your phone’s camera to track the cube as you solve it, shows the next moves on the screen, and gives voice hints along the way. @AximoLoot

Smartstudy — an AI-powered exam simulator. 🖌 The service can test your knowledge of any study material by generating questions and ready-made answers based on your uploaded content. • Supports files up to 200 MB • PDF files are supported • Generates questions tailored to your study material 💬 The program is free to use, but you’ll need your own API key. @AximoLoot

ElevenLabs just launched Eleven v4 ➡️ ElevenLabs has introduced its new Eleven v4 and v4 Turbo speech models, with a redesigned architecture focused on acting, emotions, and text-controlled voice direction. The models understand scene context, intonation, and different character roles. You can add instructions like [whispers], [laughs], or [shouts] directly into the script, along with sound events. 🔊 Voice cloning has also been upgraded. Instant Voice Clone works from a short recording, while Professional Voice Clone preserves the original speaker’s tone, delivery, volume, and accent more accurately. There are two versions: • Eleven v4 — for videos, audiobooks, ads, and characters • Eleven v4 Turbo — for voice agents and real-time applications where low latency matters ➡️ Both models are already available through ElevenCreative, ElevenAgents, and the API. @AximoLoot