Conversation Stenography uses AI to turn secret messages into harmless everyday conversations — like talking about recipes or the weather.
• Uses an LLM and a special algorithm to hide the real meaning of a conversation;
• The same AI model can decode the message when both sides know the agreed code phrase;
• Works even with lightweight models like Llama 3 4B and GPT-2, so it can run on almost any device.
@AximoLoot
Please open Telegram to view this post
VIEW IN TELEGRAM
This media is not supported in your browser
VIEW IN TELEGRAM
It promises more complex compositions, more natural-sounding instruments, and noticeably more detail across tracks.
Another major update is downloads:
• Free users get 5 lossless downloads per day;
• Pro users get up to 400 per month.
You keep the rights to tracks you create, even on the free plan. However, commercial usage depends on your subscription, and tracks using references to other artists’ songs can’t be downloaded.
@AximoLoot
Please open Telegram to view this post
VIEW IN TELEGRAM
This media is not supported in your browser
VIEW IN TELEGRAM
AirDrop for any device 🔔
Meet PairDrop — a free file-sharing service that works across Windows, Linux, Android, iPhone, Mac, and more.
@AximoLoot
Meet PairDrop — a free file-sharing service that works across Windows, Linux, Android, iPhone, Mac, and more.
• Runs directly in your browser — no apps to install;
• Open the site on both devices and select the files you want to send;
• Devices on the same Wi-Fi network find each other automatically. Different networks? Just connect using a password;
• Send files of several gigabytes without compressing or losing quality.
@AximoLoot
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
Viseron — an open-source AI-powered surveillance system for your existing cameras.
Instead of replacing your cameras with expensive “smart” models, you can add AI capabilities to the hardware you already have.
— Detects motion, people, animals and other objects, so you can focus on actual events instead of hours of empty footage;
— Supports face recognition and object detection;
— Works with cameras from different manufacturers and supports ONVIF;
— Keeps your data and recordings locally on your own hardware, with no mandatory cloud service;
— Supports a wide range of cameras and hardware acceleration options;
— Can send notifications when something important happens;
— Free to use and fully open source.
If you have a few cameras sitting around, Viseron can turn them into a much smarter surveillance system without locking you into a cloud ecosystem.
@AximoLoot
Instead of replacing your cameras with expensive “smart” models, you can add AI capabilities to the hardware you already have.
— Detects motion, people, animals and other objects, so you can focus on actual events instead of hours of empty footage;
— Supports face recognition and object detection;
— Works with cameras from different manufacturers and supports ONVIF;
— Keeps your data and recordings locally on your own hardware, with no mandatory cloud service;
— Supports a wide range of cameras and hardware acceleration options;
— Can send notifications when something important happens;
— Free to use and fully open source.
If you have a few cameras sitting around, Viseron can turn them into a much smarter surveillance system without locking you into a cloud ecosystem.
@AximoLoot
This media is not supported in your browser
VIEW IN TELEGRAM
Draw Your Font turns your handwriting into a real computer font with Claude Code — basically the tool every student once wished existed.
All you need to do is write the alphabet and numbers on a sheet of paper, take a photo, and upload it to Claude. The tool detects the characters, cleans up the background, vectorizes your handwriting, and turns it into an installable TTF/WOFF2 font. Everything runs locally, and the finished font is yours to use.
@AximoLoot
All you need to do is write the alphabet and numbers on a sheet of paper, take a photo, and upload it to Claude. The tool detects the characters, cleans up the background, vectorizes your handwriting, and turns it into an installable TTF/WOFF2 font. Everything runs locally, and the finished font is yours to use.
@AximoLoot
❤1
QuiverAI Arrow 2 — an AI tool for creating editable vector graphics.
Describe the logo or illustration you want, and the model generates a ready-to-edit SVG. You can refine the result directly in chat, adjust individual elements on the canvas, and scale the artwork to any size without losing quality before exporting it to other design tools.
For more complex compositions and advanced prompts, QuiverAI also offers a separate model called Arrow 2 Telos.
Try it here: QuiverAI
@AximoLoot
Describe the logo or illustration you want, and the model generates a ready-to-edit SVG. You can refine the result directly in chat, adjust individual elements on the canvas, and scale the artwork to any size without losing quality before exporting it to other design tools.
For more complex compositions and advanced prompts, QuiverAI also offers a separate model called Arrow 2 Telos.
Try it here: QuiverAI
@AximoLoot
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
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 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
Please open Telegram to view this post
VIEW IN TELEGRAM
Please open Telegram to view this post
VIEW IN TELEGRAM
This media is not supported in your browser
VIEW IN TELEGRAM
A Chinese developer is making a game where you have to sneak out of work early without getting caught by your boss.
@AximoLoot
@AximoLoot
Grok 4.7 is out — Musk’s most powerful AI model to date.
xAI has upgraded its flagship model for coding, documents, and long-running agentic tasks.
What’s new:
• Handles long context better;
• Checks its own work more carefully;
• Performs better on engineering tasks;
• Natively understands the Grok Bot environment.
API pricing remains the same as Grok 4.6 — $2 per million input tokens and $6 per million output tokens.
You can already try it here: Grok Build
@AximoLoot
xAI has upgraded its flagship model for coding, documents, and long-running agentic tasks.
What’s new:
• Handles long context better;
• Checks its own work more carefully;
• Performs better on engineering tasks;
• Natively understands the Grok Bot environment.
API pricing remains the same as Grok 4.6 — $2 per million input tokens and $6 per million output tokens.
You can already try it here: Grok Build
@AximoLoot
This media is not supported in your browser
VIEW IN TELEGRAM
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.
➡️ Just open the tab, drop in your file, and upscale it. No installation required.
@AximoLoot
• 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.
@AximoLoot
Please open Telegram to view this post
VIEW IN TELEGRAM
❤1
Media is too big
VIEW IN TELEGRAM
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.
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 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.
@AximoLoot
Please open Telegram to view this post
VIEW IN TELEGRAM
🥱1