Vibe Coding: Build Apps with Claude Code, Cursor, and Codex

Install essential tools by hand to understand the setup before using coding agents like Codex, Cloud Code, Open Code, or the Gemini CLI. Ensure Node.js is installed for some agents.

Install and verify the full tech stack—python, javascript, typescript, node.js, react, next.js, tailwind, git, nvm, pyenv, pip, uv, bun—and learn what each tool does and how to check your setup.

Install python, pip, and uv with the simpler method from python.org, and consider pyenv for version management across windows, mac, and linux.

Install node.js and npm, then optionally set up nvm for version management or install bun; verify node and npm versions and prepare to run servers with npm run dev.

Learn the differences between Git and GitHub, and how Git provides local version control with branches and version history while GitHub hosts remote repositories and enables publishing and collaboration.

Install Python, pip, uv, pyenv, Node.js, npm, and nvm to prepare your environment, and apply the idea that you only learn if you do with git and GitHub.

Explore ai in the terminal with cli tools like Gemini CLI, OpenCode, CloudCode, and Codex, and master Git and GitHub for version control.

Compare instant builders, AI native IDEs, and autonomous CLI tools for wipe coding. Leverage tools like lovablebolt.new, base44, horizons, Cursor, Copilot, Cloud Code, and OpenCode to prototype fast.

Explore the Gemini CLI in the terminal, its free tier, Google authentication, install options, and slash commands for file operations, web fetching, and Gemini 3 models.

Navigate a folder with the Gemini CLI and view results in VS Code; Gemini writes files, runs syntax checks, and creates gemini.md to share project context with coding agents.

Learn to run multiple coding agents in one project by installing and using OpenCode, Codex, and Claude Code, and switching between Gemini, Cloud Code, and codecs.

Master git and github for version control, including commits, push, pull, add, clone, and checkout. Understand git ignore, staging area, and rolling back to older commits.

Discover how coding agents in cloud code use plan mode to autonomously draft a bodybuilding nutrition book, structure chapters, and push updates via GitHub with version control.

Recap the core tools for rapid MVP creation—instant builders, AI native IDs, and powerful CLI and coding agents—while highlighting Git for version control and treating nearly everything as code.

Engineer context window with cloud code, codex, gemini cli, or open code. Build a local financial app that tracks income, expenses, and stock performance, with branding and monetization.

Explore context engineering fundamentals, from selecting the right model and defining the tech stack to configuring environments and prompts for efficient LLM workflows.

Set up cloud code to build finance app in VS Code, configure a status line for token and context tracking, and practice context engineering with Opus 4.6 and slash commands.

Build a finance tracking app that records income, expenses, and assets, tracks a stock portfolio with up-to-date prices, using Next.js, TypeScript, Tailwind CSS, SQLite, Prisma, Yahoo Finance, and Lucid React.

Test and debug the budgeting app, fix dark mode, and push updates to GitHub, while keeping all data local and tracking transactions and a portfolio with Cloud Code.

Build a virtual try-on app with user uploads, item selection, and generate-try-on, backed by a monitorable nlm workflow, with Next.js frontend, Superbase authentication, Stripe payments, and Vercel hosting.

Create the app backend with n8n, implement a webhook-based workflow for image editing, monitor executions, publish v1 and v2 workflows, and ensure observability and authentication for production.

Develop the frontend with cloud code, building a Next.js 16 app in TypeScript, styled with Dalvin CSS, and integrating login via Supabase, payments with Stripe, and hosting on Vercel.

Push your app to GitHub, link the remote repository to Vercel, and deploy automatically to a public URL. Configure environment variables and webhook URLs, then redeploy to apply changes.

Learn to build a login with Supabase using the model context protocol (MCP) to wrap API access, connect via cloud code, and manage authentication from setup to deployment.

Implement a Stripe-based payment flow for a try-on app, using MCB cloud code, Playwright testing, and Superbase to manage a 9.99 monthly subscription with a payment link and webhook synchronization.

Automatically monitor n8n workflows in production with the error trigger and Gmail alerts, publish versions, and receive actionable emails when a workflow breaks to enable quick fixes.

Integrate a chatbot and end-to-end styling workflow, manage product images, test locally, then push to production on Vercel, and monitor, secure, and market your fashion app.

Build a fully local image generation app using ConvUI as backend, with setup, frontend via Cloud Code, and a two-image, prompt-driven workflow for virtual try-on.

Claude Fable 5 and Sonnet 5 models appear in cloud and terminal; Fable 5 is strongest but time-limited and API-priced, so use Opus for most tasks and avoid Sonnet 5.

Highlight context engineering as the foundation for building apps. It covers choosing the right model, defining tech stack, and setting up environment with CLI tools and VS Code.

Explore advanced Cloud Code workflows, from permission hacks and custom slash commands to multi-agent systems, skills, hooks, and open-source LLMs, building interactive apps with Gemini embeddings and self-improving AI.

Master access permissions in cloud code by using slash permissions to manage allow, deny, and workspace rules, and learn why dangerously-skip-permissions can bypass checks.

Explore how to create custom slash commands inside the .clod setup, using a clod.md and a commands folder to save repetitive prompts like git commit or git push.

Explore running multiple agents and sub-agents in parallel to accelerate backend and frontend work, saving tokens, using cloud code and codex while preserving separate context windows.

Master quick prompting and negative prompting by crafting a concise cloth.md with restrictions that guide cloud code, emmet, agent memory, agents, and custom hooks.

Learn to implement hooks in Cloud Code, using pre-tool and post-tool events to automate actions, logging, and notifications, guided by the hook lifecycle and practical JSON examples.

Explore creating and using agent skills with cloud code, including progressive disclosure, open standards, and external skill hubs like skills.sh, while managing context window and token usage.

Install and integrate skills from skills.sh or the GitHub repo using the skill finder and npx. Build web pages with front-end design skills across cloud code, cursor, codex, and gemini-cli.

Explore using Remotion inside Claude Code to create videos and animations programmatically with Cloud Code, including stock chart and logo fight b-roll rendered via FFmpeg and assets.

Explore Anthropic's 32-page skill guide and GitHub repo, learn with the skill creator via Cloud Code, and see how OpenClaw docs can inform skill improvements.

Explore auto research from Andrej Karpathy, an open source self-improving ai loop that autonomously edits code, trains briefly, and improves skills like OpenCloud docs.

Learn to use and build plugins that bundle agents, hooks, mcp servers, and slash commands to share across teams via a plugin marketplace, or use standalone configurations for personal workflows.

Discover how to empower AI to test itself with strict test driven development, user stories and acceptance criteria, prompts, and plugins, achieving 80 percent test coverage and self-improvement.

Explore how CLIs like GitHub CLI, Google CLI, and Playwright CLI boost coding agent workflows with lean, token-efficient commands, plus tips to install, authenticate, and use CLIs alongside MCB.

Build a multimodal rag application using Gemini embeddings integrated with a vector database, enabling text, images, videos, audio, and PDFs to be embedded, stored, and queried.

Build cloud agents with Anthropic's claude-agent-sdk in Python, creating agent loops that gather context, take actions, and verify outcomes. Explore context maintenance, long-running agents, and semantic search.

Discover how to run coding tasks entirely offline with Ollama, selecting local LLMs from QUEN 3.5, GLM, and GPT OSS families, and using Cloud Code to code locally.

Ralf loops and cloud code enable autonomous, days-long tasks by looping prompts, preserving memory in files and git history, and scheduling cron-like recurring actions.

Master voice mode and slash commands in cloud code to speed prompt writing, manage sessions and branches, and tune memory and configs with practical tips.

Explore how Boris leverages Cloud Code to run multiple parallel instances, plan-driven workflows, slash commands, and dedicated sub-agents, with a continuous verification loop to boost code quality.

Master the recap by detailing access permissions, slash commands, multi-agent work in parallel, hooks, skills, plugins, verification, command line interfaces, the agent SDK, Gemini embeddings, cron jobs, and voice mode.

Learn to build a simple chatbot MVP using Python and an OpenAI-backed AI agent with Pydantic AI, using Cursor and Gradio for the frontend.

Discover how to install and use cursor, a vscode fork, to edit projects, create files with an agent and plan mode, review diffs, and run python snake games with pygame.

Build an autonomous research agent stack with Python, Pydantic AI, Gradio, and Cursor. Orchestrate multi-agent workflows for real-time web research, scraping, and parallel execution using OpenAI models.

Build a simple chatbot prototype with Pydantic AI and Gradio, using a GPT-5 mini model and the llms.txt documentation to chat with an AI agent.

Tag and load external documentation in Cursor to provide project-wide context, using docs, indexing, and entry-point prefixes, then search and browse pedantic docs to build a pedantic ai agent.

Develop a deep research agent with a Gradio frontend in Python using Pydantic AI, orchestrating multi-agent web research (DuckDuckGo) and producing structured reports with executive summaries and evidence.

Explains the multi-agent pydantic ai project, detailing environment setup and api key security. Outlines the architecture with classification, planning, research, and writer agents, plus gradio front-end and markdown reporting.

Push your Gradio app to GitHub and share a public one week URL with friends, then explore hosting options like Hugging Face or Vercel while safeguarding API keys.

Explore PydanticAI agents in Python by leveraging GitHub examples and docs, and apply evaluations and observability with logfire, avails, and vector databases to tailor your apps.

Recap the core tech stack—Cursor, Python, Pip, Python DKI, Gradio—and how to build, publish on GitHub, and share chatbots and projects like Deep Research.

Install antigravity, explore its familiar interface alongside cursor and VS Code, and learn to use built-in models such as Gemini 3.1 pro and cloud code for development.

Build a content repurposing app that turns a YouTube transcript into posts for X, Instagram, and LinkedIn, with AI-generated images via Gemini and NanoBanana, using a React and Vite frontend.

OpenClaw is an open-source personal assistant that orchestrates LLMs, with skills, plugins, and cron jobs. Install it locally or on a VPS and connect via API keys or Telegram.

OpenClaw is a self-hosted, open-source ai agent that acts as an orchestrator, connecting chat apps to a proactive llm brain, with customizable sub-agents, cron triggers, and self-modifying capabilities.

Install OpenClaw across Linux, macOS, Windows or a VPS, choose secure deployment options, follow onboarding, configure AI providers and tokens, and connect via Telegram for hands-on workflow.

Learn to set up OpenClaw's cloud bot, connect it to Telegram, train its persona, manage tokens and costs, and explore agents, skills, and config for a personal AI assistant.

OpenClaw is an open source personal AI assistant that runs on your machine or a VPS, with a one-liner installation and bootstrap onboarding, connecting multiple models.

Explore voice agents and AI avatars built with python, livekit, cursor, and codex; learn how to deploy a customer support agent across web and phone with vector databases.

Explore the goal, strengths, weaknesses, and tech stack of AI voice agents, including speech-to-text, llms, and text-to-speech, plus cost and deployment considerations.

Explore a simple GitHub project that builds a real-time ai avatar with LiveKit, agentworker.py, and OpenAI real-time mode using the Aloy voice and Tavos persona.

Explore LiveKit, a platform for building ai voice apps and avatars. This overview covers the interface, docs, quickstarts, and sample repos to get you started with python or node.js.

Explore livekit driven Python voice agents with locally run AI models, ready-made examples, and code you can customize for avatar agents, lip sync, and deployment.

Recap the basics of voice AI agents, transcription via text-to-speech, LLM processing, function calls, and multi-modal options, with LiveKit and hosting tips.

Tackle compliance, law, and security risks essential for production-ready apps, including jailbreaks, prompt injections, data poisoning, authentication, API keys, privacy, and open-source licenses.

Examine jailbreaks in AI systems, from many-shot and zero-shot prompts to guardrails and frontier models, with real-world demos and potential risks for applications.

Explore prompt injections that threaten internet-enabled coding agents like OpenCLoud, Cloud Code, and Cursor, risking exposure of emails and personal data through malicious prompts.

Explore data poisoning and backdoor attacks in open-source fine-tuned models. Identify how instruction training and specific words can influence model behavior, and examine data security and privacy concerns.

Explore the Model Context Protocol (MCP) and how to implement MCP servers in cloud code, codex, cursor, n8n, or python, while recognizing security risks and the dangers of unsafe servers.

Protect against tool poisoning, MCP rug pulls, and shadow tool descriptions by understanding how malicious MCP servers can exfiltrate data and alter tool behavior, and adopting robust cross-server security.

Never expose your api keys; store them securely in a .env file, rotate them regularly, and use authentication for servers to stay safe when publishing or sharing code.

Understand how copyrights affect creating, selling, and publishing AI outputs—from chatbots and stories to images and voices—and OpenAI's copyright shields and training data considerations.

Explore open source licenses, including MIT and Apache 2.0, plus GPL, AGPL, fair code, and custom licenses, with guidance on commercial use and restrictions.

Explore data protection for client and private data when using ai agents, including ownership of inputs and outputs, retention and access, and security of OpenAI, Grok, and Gemini APIs.

Explore censorship, alignment, and bias in AI agents, comparing DeepSeek and OpenAI restrictions with uncensored Dolphin models that offer steerable alignment, local data control, and ethical guidelines.

Explore the EU AI Act and GDPR requirements for chatbot apps, mapping risk levels from unacceptable to minimal and detailing compliance steps, data privacy, consent management, and bias mitigation.

Examine jailbreaks, prompt injections, data poisoning, and MCP server risks, and implement safe api key handling, license awareness, and compliant, secure ai app development.