分类筛选
找到 1605 个 Demo
Vibe-coded an experimental hero header. Midjourney...
Vibe-coded an experimental hero header. Midjourney → GPT Image 2 → Google Ai Studio → Codex. Logo built in QuiverAI as native SVG. Full workflow breakdown in replies. Live: https://t.co/0fiwt6kM6m Code: https://t.co/1JQlPZTGeG https://t.co/9JMapjmZ13
This "Image to Code" (built with Gemini 2.0) is in...
This "Image to Code" (built with Gemini 2.0) is insane 😆 https://t.co/YepOynYlVJ
This is literally my new workflow now: Real-time...
This is literally my new workflow now: Real-time research/search → Grok 4 Planning & Reasoning → Grok 4 Heavy Coding → Claude 4 Sonnet w/ Claude Code Write Test Cases → Gemini 2.5 Pro Run Test Cases → Codex Debug → o3 Bookmark this.
I built a Street Fighter style game on using Vib...
I built a Street Fighter style game on @base using Vibe Coding and a proper dev stack. I started with Claude but switched to Google AI Studio for the heavy lifting. I pushed through multiple rounds of fixes to polish the experience: - Smooth KO animations - Responsive mobile friendly controls - Smarter AI attack patterns Then I took it to production: -Code managed on GitHub - Frontend deployed on Vercel - Smart Contract live via Remix on Base Wallet Connect integrated for the onchain leaderboard The game is playable with or without a wallet. Connect to submit your score onchain and compete globally. It was not a one prompt miracle but the speed is still mind blowing. Mobile gameplay below 🕹️👇
Deepclaude is an OpenAI-compatible API that chains...
Deepclaude is an OpenAI-compatible API that chains DeepSeek R1 with Claude or Gemini models for improved reasoning and output quality. - Supports programming workflows with DeepSeek R1 + Claude Sonnet 4 - Enables content creation with DeepSeek R1 + Gemini 2.5 Flash or Pro - Provides a graphical configuration interface for simplified deployment - Compatible with OpenAI format for easy integration into existing tools Explore it here: https://t.co/4hxnERk1RR
تعبت من التنقل بين أدوات الـ SEO، تحليل المنافسين،...
تعبت من التنقل بين أدوات الـ SEO، تحليل المنافسين، والنسخ واللصق بكل مقال تكتبه؟ 🤯 عندي لك اليوم 'Keywordo-kun'، وكيل ذكاء اصطناعي راح يختصر عليك كل هذي الدوخة ويجهز لك مقالات متصدرة بـ دقائق!" العملية التقليدية للـ SEO تاخذ وقت كلش طويل، بس هذا المطور كدر يجمع كل مراحل الـ SEO (بحث كلمات مفتاحية، تحليل منافسين، وكتابة محتوى) بـ "وكيل ذكاء اصطناعي" واحد باستخدام n8n. ليش هذا الـ SEO Agent يفرق عن غيره؟ (القيمة المضافة): 1- واجهة تفاعلية: تحجي وياه مثل ما تحجي ويا مساعدك الشخصي، وهو ينفذ. 2- بيانات حقيقية: يستخدم أدوات مثل DataForSEO و Firecrawl حتى يجيب بيانات حقيقية من الـ SERP (نتائج البحث)، يعني المقال اللي ينكتب مو 'تخمين' من الـ AI، هو مبني على بيانات واقعية. 3- هيكلية ذكية: المطور استخدم نظام (العقل + الأدوات)، حيث يفصل الـ Agent (Gemini 2.5) عن الأدوات، وهذا يخلي النظام مستقر وسهل التطوير. 4- كتابة احترافية: ياخذ كل النتائج ويحولها لمقال طويل، حصري، ومحسن لمحركات البحث. شلون يشتغل تقنياً؟ 1- الواجهة: تستخدم v0 (من Vercel) كواجهة شات. 2- العقل: محرك n8n مع Google Gemini 2.5 Pro. 3- الأدوات: سير عمل فرعي (Sub-workflow) متصل بـ APIs لجلب بيانات SEO دقيقة. الروابط: 🌐 جرب الـ Live Chat: https://t.co/7irYkWP12C 🎥 فيديو الشرح بالتفصيل: https://t.co/RuqN9KOdTA هذا المشروع هو مثال ممتاز لكيفية استخدام الأتمتة (n8n) مو بس للترفيه، بل لتحويل عمليات البيزنس المعقدة إلى عملية بضغطة زر. سؤالي إلك: شنو أكثر خطوة تزعجك بعملية كتابة المحتوى؟ وهل تشوف إن الـ AI كدر يحلها فعلاً؟
Gemini CLI is our new #OpenSource AI agent that br...
Gemini CLI is our new #OpenSource AI agent that brings the power of Gemini directly into your terminal! Access Gemini 2.5 Pro with 1M token context window, 60 requests/min, and 1,000 requests/day—at no cost with a free Gemini Code Assist license → https://t.co/Fp0wdxfbEW https://t.co/Ul6hjn1hKQ
"#Google vient de présenter une nouvelle #IA appel...
"#Google vient de présenter une nouvelle #IA appelée #Gemini3.5 Live Translate, capable de générer des traductions instantanées avec seulement quelques secondes de retard, par rapport à l’orateur"💡😳 https://t.co/JRRc6MZS65 via @pressecitron
今日のチーム練習❗️ 1000+3000+1000レスト3分 来週のMK出れないけどどこかで5000...
今日のチーム練習❗️ 1000+3000+1000レスト3分 来週のMK出れないけどどこかで5000のレースにはチャレンジしたい #まるお製作所RC #Gemini3 #ジェミラン https://t.co/Nitb0CIkO4
GOOGLE AI STUDIO JUST MADE APP BUILDING WAY TOO EA...
GOOGLE AI STUDIO JUST MADE APP BUILDING WAY TOO EASY You can now build a real Android app from a browser prompt, but most people will miss the business use case. What Changed: → Open Google AI Studio in your browser → Go to the Build tab → Pick Build an Android app → Type what you want → Watch it generate real native Android code Why This Matters: ✓ No heavy software setup ✓ No coding required ✓ No downloads to start ✓ Apps can use GPS, Bluetooth, camera, and more ✓ Built-in Android phone simulator lets you test the app on screen Business Apps You Can Build: → Lead capture app for events → Community signup app → Weekly content planner → Member welcome app → Coaching call schedule app The real win: Google AI Studio now lets simple business ideas become usable phone apps in minutes instead of weeks. The people who learn this early will build internal tools faster than teams still waiting on developers.
There is a GitHub repo that lets you code with any...
There is a GitHub repo that lets you code with any AI model you want. OpenAI. Anthropic. Google. Meta. DeepSeek. Mistral. Any of them. All of them. Free. Open source. No monthly fee. It is called OpenCode. And here is everything you need to know. OpenCode is the most-starred open-source agentic coding tool in 2026 over 150,000 GitHub stars, 850+ contributors, 11,000+ commits, and roughly 6.5 million developers using it monthly. It does what Claude Code does, read your codebase, edit files, run commands, manage git, chain multi-step tasks but lets you plug in any model from 75+ providers, including fully local inference via Ollama at zero API cost. Here is why that matters. Every major AI company wants you locked into their model. OpenAI wants you on GPT-5. Anthropic wants you on Claude. Google wants you on Gemini. Each one charges a separate subscription. Each one has a separate interface. Each one is a walled garden designed to keep you inside it. OpenCode has no walls. GPT-5.4 for complex reasoning. Claude Sonnet 4.6 for coding. Gemini 3 Pro for long context. DeepSeek V4 for cost efficiency. Llama 4 locally via Ollama for complete privacy. Switch between all of them. In the same tool. Without changing anything except the model name. Most developers running OpenCode with Claude Sonnet 4.5 via direct API pay around six dollars per day on heavy use, a fraction of the subscription cost of any single AI coding tool. Six dollars per day on heavy use. No subscription. No vendor lock-in. Here is what OpenCode actually does that makes it a genuine Claude Code alternative. OpenCode spawns Language Server Protocol servers and feeds compiler diagnostics back to the model after every edit. If the agent introduces a TypeScript type error, the next round includes the error and the model self-corrects. This is unique to OpenCode in 2026. Claude Code does not feed LSP diagnostics into the loop by default. The agent catches its own type errors. Automatically. Before you see them. Claude Code does not do this. Here is the full feature comparison nobody is posting clearly enough. OpenCode vs every paid coding agent: 🤖 Models → OpenCode: 75+ providers including local. Others: one model each. 💰 Price → OpenCode: free as a tool. Others: $20 to $200 per month. 🖥️ Desktop app → OpenCode: included. Claude Code: terminal only. 🔒 Local inference → OpenCode: via Ollama, zero cost. Others: cloud only. 🔧 LSP integration → OpenCode: built in. Claude Code: not available. ⭐ GitHub stars → OpenCode: 150,000. Claude Code: 71,500. 🌐 License → OpenCode: MIT. Others: closed source. Here is the story that explains the 150,000 stars. In January 2026, Anthropic blocked OpenCode from using Claude via OAuth login, an attempt to protect their market. OpenCode doubled its stars in two months. The relationship between them is now openly adversarial. Anthropic tried to slow it down. It made OpenCode more popular. Here is the honest part because every comparison deserves one. Claude Code dominates actual usage, 4% of all public GitHub commits, 135,000 per day. Claude Code is a vertical integration masterpiece. OpenCode is a horizontal flexibility play. Claude Code is more polished. Its agentic loop is tighter. Its SWE-bench performance still leads on the hardest real-world coding benchmarks. But here is the question every developer paying $200 a month should ask themselves. Do you need the best single model or the freedom to use the right model for each task? If the answer is freedom, OpenCode is already on GitHub. Already free. Already supporting every model you use. The AI industry spent years building walls between models. OpenCode knocked them all down. Source: OpenCode · All Hands AI · Morph LLM · OpenAI Tools Hub · May 2026 (Link in the comments)
毎週水曜日の #Gemini3 練習会 メニュー： ・分割走（1000 - 3000 - 1000m...
毎週水曜日の #Gemini3 練習会 メニュー： ・分割走（1000 - 3000 - 1000m） 走る前はどうかなぁ、という感じだったけど、概ねイメージ通りのペースで走れた これを分割ではなく、通しでちゃんと走れるようになりたい #GeminiRunners #まるお製作所RC #ジェミラン https://t.co/fx11ueTWCm