mozgi
SkillsCatalog of skills for Claude Code and other AI agents
MCP serversTop MCP servers from GitHub: browser, Figma, databases, docs
Online whiteboardDiagrams, ideas and moodboards — no sign-up.
Local AI chatA language model right in your browser on WebGPU — private,
AI text checkerCheck whether a text was written by AI — right in your brows
AI prompt libraryImage and video prompts with examples and model filters.
Video prompt builderA prompt for Seedance, Kling, Veo, FLUX — in clicks.
Image compressorResize and convert images to WebP, JPEG, or PNG.
Crop imageCrop by frame with aspect ratios and exact pixel size.
AI image upscalerEnlarge images 2× or 4× with ESRGAN.
GIF StudioGIFs from video and photos, compress, add text.
SVG to 3DTurn SVG artwork into 3D shapes with adjustable materials an
Depth map generatorCreate depth maps and contour maps from photos and videos.

Prompts for images and video

Prompt builder

Everything for AI agents

  • Skills for Claude Code, Codex and Cursor
  • MCP servers with setup instructions
  • Local AI chat and AI text checker
Online whiteboard

One board. All your ideas.

Diagrams, mind maps, flowcharts and sticky notes: draw processes and plans on an infinite board right in your browser

Open the board

Image processing

3D

From flat to 3D

Turn an SVG logo or icon into a 3D model

Open SVG to 3D

News

Fresh from the world of AI: models, services and updates

Sandflow has released the Sand3D dataset of 106 thousand 3D models

Sandflow has released the Sand3D dataset of 106 thousand 3D models

French startup Sandflow has released an open dataset called Sand3D, consisting of 106 thousand 3D models of various objects. The dataset is designed for training generative models to create and edit 3D graphics. Each of…

FAA expands use of AI for air traffic control in the US

FAA expands use of AI for air traffic control in the US

The US Federal Aviation Administration (FAA) has announced the expansion of AI tools to optimize air traffic control. The system, previously tested in select areas, now covers the entire national airspace, helping contr…

Nebius to invest $40 billion in expanding AI infrastructure

Nebius to invest $40 billion in expanding AI infrastructure

Nebius has announced plans to invest $40 billion in developing its own artificial intelligence infrastructure. The large-scale project will be funded through customer prepayments and debt financing. The new "neocloud" f…

Brave Search API has added payments for AI agents via x402

Brave Search API has added payments for AI agents via x402

Brave has added support for real-time micropayments for autonomous AI agents. The update to the search interface integration is implemented based on the open x402 protocol. The new mechanism allows for on-demand payment…

Apple introduces Foundation Models Framework for macOS

Apple introduces Foundation Models Framework for macOS

Foundation Models Framework is a new native Swift API designed to integrate large language models directly into macOS applications. The tool allows developers to implement AI capabilities with support for local on-devic…

Cornell University study: data centers shift infrastructure costs onto residents

Cornell University study: data centers shift infrastructure costs onto residents

Researchers at Cornell University have published a report on the financial burden that energy-intensive facilities place on regions. The analysis showed that AI data centers cover only 38% of the costs for expanding loc…

Chinese AI models reveal traces of sanctioned Nvidia accelerators

Chinese AI models reveal traces of sanctioned Nvidia accelerators

Forensic analysis of advanced language models from China has uncovered hardware traces of Nvidia GPUs subject to export restrictions. The study covers products from developers such as DeepSeek, Moonshot AI, MiniMax, and…

The Muse service has added an automatic calling feature for U.S. businesses

The Muse service has added an automatic calling feature for U.S. businesses

The Muse platform has introduced a new capability for its users: an AI agent can now make outgoing phone calls to commercial organizations within the United States. Previously, the service's functionality was limited to…

Google has introduced the flagship Gemini 4 Hydrogen model

Google has introduced the flagship Gemini 4 Hydrogen model

Google DeepMind has announced the release of Gemini 4 Hydrogen, a new flagship model focused on high-performance computing and complex logical tasks. The developers state that the model's architecture has been completel…

The US Congress has launched an investigation into Nvidia's activities

The US Congress has launched an investigation into Nvidia's activities

Democrats in the US Congress have initiated an official investigation into Nvidia. Lawmakers sent a scathing letter accusing the company's leadership of implementing questionable schemes in their interactions with the D…

Anthropic adds Claude to Google Meet video calls

Anthropic adds Claude to Google Meet video calls

Anthropic has introduced Claude Meetings, a tool designed to integrate artificial intelligence into video conferencing. The new feature allows the assistant to join sessions in Google Meet. During a video call, the syst…

SoftBank lost $3 billion on turbine purchases for AI data centers

SoftBank lost $3 billion on turbine purchases for AI data centers

Japanese conglomerate SoftBank has reported losses exceeding $3 billion resulting from advance orders for gas turbines. The equipment was intended to provide power for future data centers focused on artificial intellige…

What is mozgi.io

mozgi.io is a site for people who create images and video with AI. No sign-up required.

Prompts and the builder

The catalog has more than 33,000 ready-made prompts for GPT Image, Nano Banana Pro, Seedance, Grok Imagine Video and other models. Each one comes with a sample result, the author and the source. For video there is a builder: choose a genre, a character, the camera, lighting and sound, and it will put together a prompt for Seedance, Kling, Veo and other video models.

Browser tools

Photo compression to WebP, JPEG and PNG, cropping to 1:1, 16:9 and 9:16, 2x and 4x upscaling, GIFs from video and photos, depth maps, 3D models from SVG and a local AI chat. Files are processed on your device and are not uploaded to a server. AI-powered tools download the model into your browser on first launch, and the local chat needs a browser with WebGPU support.

Skills and news

More than 800 skills for Claude Code and other AI agents: documents, code, design, content. Each skill has a description and a link to its source. The news section has short notes about models and AI tools in five languages.

Who it's for

Designers, social media managers, marketers and online sellers, content creators, and developers who work with AI agents.

Privacy

The site uses visit analytics: Cloudflare Web Analytics without cookies, while Yandex Metrica and Google Analytics are enabled only after you give consent. Learn more in the privacy policy.

Frequently asked questions

How much does mozgi.io cost?

The prompt catalog, news, skills and some of the tools are open with no sign-up. The rest is available with a subscription: mozgi.io original prompts, prompt translations, search, audience collections and the locked tools (compression, upscaler, GIF studio, video prompt builder, local AI chat, local AI text detector, online whiteboard). You can try the locked tools for one hour. Generating images and video is paid separately, in the AI services themselves.

Are my photos and videos uploaded to a server?

No. Compression, cropping, upscaling, the GIF studio, depth maps, SVG to 3D and the local AI chat process your files and conversations in your browser. Analytics counts page visits but does not see your files.

Does mozgi.io generate images and video?

No. Here you'll find prompts and tools for preparing and processing files. You run the generation itself in an AI service: GPT Image, Nano Banana, Seedance and others.

What languages is the site available in?

Russian, English, Turkish, Chinese (Traditional) and Italian.