Key terms and concepts in AI-powered application development. From vibe coding to code healing — understand the tools and techniques shaping modern development.
An autonomous AI building mode that breaks complex app requests into 5-15 sequential tasks, executing each with error recovery and checkpoints. The AI plans the implementation, executes tasks one by one, retries failures up to 3 times, and allows users to pause, resume, or stop at any point.
A platform that uses artificial intelligence to generate functional web applications from natural language descriptions. Unlike traditional no-code tools that use drag-and-drop, AI app builders produce real source code that users own and can export. Examples include GenMB, Bolt.new, and Lovable.
The process of using large language models (LLMs) to produce source code from natural language prompts. Modern AI code generators like GenMB produce complete, multi-file applications rather than just snippets, using pipeline stages for validation, generation, error healing, and enhancement.
A set of rules and protocols that allow different software applications to communicate with each other. APIs define endpoints, request formats, and response structures. GenMB generates REST API endpoints as part of full-stack applications and connects to third-party APIs through its plugin system (Stripe, Supabase, Twilio, etc.).
Zero-configuration usage tracking automatically injected into every deployed GenMB app. Tracks page views, unique visitors, and user sessions without any setup. View metrics in the app's Analytics tab. Available on Business plan.
A distributed network of servers that delivers web content to users from the closest geographic location. GenMB uses esm.sh as a CDN for JavaScript module imports, allowing generated apps to load dependencies instantly without npm or build tools.
A browser extension for GenMB that captures any webpage as a screenshot and sends it to GenMB as a starting prompt. Point at a design you like — a competitor's page, a UI inspiration — and GenMB generates a similar app. Useful for "browse-to-app" workflows without manual design description.
The ability to download a complete GenMB project as a ZIP archive containing all source files, assets, configuration, and dependencies. Exported code is plain HTML/CSS/JS, React, or React+TypeScript with no GenMB-specific bindings — deploy it anywhere: Vercel, Netlify, your own server, or a client's hosting.
GenMB's automatic error detection and self-fixing system integrated into the code generation pipeline. When AI-generated code has issues (syntax errors, missing imports, broken JSX, logic errors), Code Healer identifies them and iteratively fixes them using tool-based editing with JSON fallback, up to 3 times per generation.
A collection of pre-built, reusable UI elements (buttons, modals, forms, cards) that follow consistent design patterns. GenMB uses shadcn/ui components internally and generates apps using popular component patterns. Component libraries accelerate development by providing tested, accessible building blocks instead of coding each element from scratch.
A user-owned domain name (e.g., myapp.com) connected to a deployed application instead of the default subdomain (myapp.genmb.com). GenMB Pro users can connect up to 5 custom domains with automatic SSL certificate provisioning.
An integration that connects a GenMB app to an external data source. GenMB supports four data connectors: Airtable, Google Sheets, Notion, and REST APIs. Configure a connector in the Integrations panel, and the AI generates code that reads, writes, and syncs data between your app and the external source, handling authentication and pagination automatically.
The process of making an application accessible on the internet. GenMB offers multiple deployment options: subdomain deployment (yourapp.genmb.com), custom domains with SSL, shared Cloud Run backend hosting, and dedicated GKE Autopilot pods with pause/resume for cost control.
A visual theme editor inside the GenMB app editor that extracts design tokens (colors, fonts, spacing) from the live preview. Users can swap colors, change typography, apply pre-built theme presets, or create custom themes — with instant live preview. When applied, the AI rewrites the source code with the new design system rather than injecting inline styles.
A platform for packaging applications into isolated containers that include all dependencies, configuration, and runtime. Containers ensure an app runs identically on any machine. GenMB uses Docker containers for its runtime sandbox (isolated code execution), validator service (Playwright-based testing), and deployment infrastructure on Cloud Run and GKE.
A JWT-based access credential that allows a deployed GenMB app to be embedded as an iframe on an external website. Embed tokens specify which domains are authorized to embed the app, providing access control for white-label or embedded app use cases. Generate tokens in the deployment settings.
The official JavaScript module system standardized in ECMAScript 2015. GenMB uses ESM imports via the esm.sh CDN, allowing generated apps to import libraries like React and Chart.js directly in the browser without npm, webpack, or any build step. This enables instant preview and simpler deployment.
AI-powered chatbots built in GenMB that respond to messages on Telegram, Slack, and Email. Each assistant has a configurable personality, knowledge base, and set of scheduled tasks. Assistants are deployed as dedicated GKE pods and can hold multi-turn conversations, answer questions from uploaded documents, and trigger actions on a schedule. Available on Business plan.
A GenMB Pro feature that keeps your app's code synchronized with a GitHub repository. Changes made in the GenMB editor automatically push to the linked repo. You can also pull changes from GitHub back into GenMB, enabling hybrid workflows where AI generation and traditional development happen in parallel. Supports public and private repositories.
A query language for APIs developed by Meta that lets clients request exactly the data they need in a single request, unlike REST where each endpoint returns a fixed structure. GraphQL reduces over-fetching and under-fetching of data, making it efficient for complex frontends that display data from multiple sources.
A content management system that stores and delivers content through an API without a built-in frontend. Unlike WordPress (which couples content and presentation), a headless CMS lets developers use any frontend framework. Examples include Strapi, Sanity, and Contentful. GenMB can generate apps that consume headless CMS APIs through its data connector system.
A development feature that instantly reflects code changes in the running application without a full page refresh, preserving application state. GenMB's live preview updates automatically after each generation or edit, giving immediate visual feedback without manual browser refreshes.
The GenMB community hub for sharing and discovering apps. Users publish their apps as templates that others can browse, preview, rate, and install as starting points for their own projects. The Marketplace tracks install counts and ratings. Publishing is free; any GenMB user can contribute.
An open standard by Anthropic for connecting AI tools to external services via a standardized protocol. GenMB exposes an MCP server with 7 tools (list apps, get app details, generate code, refine code, deploy, get deployment status, list templates) accessible to Claude Code, Cursor, and other MCP-compatible clients. Authenticate with a GenMB API key (genmb_sk_ prefix) via Streamable HTTP.
An architecture where an application is split into small, independent services that communicate over APIs. Each service handles one domain (auth, payments, notifications) and can be deployed and scaled independently. Contrast with monolithic architecture, where all functionality lives in a single codebase.
The ability to generate applications in multiple JavaScript frameworks from the same platform. GenMB supports 3 frameworks: Vanilla JS (HTML/CSS/JS), React, and React + TypeScript, allowing users to choose the best framework for each project.
A GenMB feature where the AI creates a structured breakdown of what it will build before generating code. Plan Mode produces higher quality output for complex, multi-feature applications by thinking through components, data models, UI layout, and technical trade-offs first.
GenMB's integration framework that injects service-specific context into AI prompts during code generation. Unlike post-hoc integrations, plugins (Supabase, Stripe, Clerk, etc.) make the AI natively understand how to use each service. Users enable plugins and provide API keys; the AI generates properly integrated code.
The practice of crafting effective natural language descriptions to get better output from AI systems. In GenMB, good prompts specify the app type, key features, data structure, and user flow. More specific prompts produce better applications. Plan Mode and Knowledge Files complement prompt engineering.
A web application that uses modern browser features to deliver an app-like experience, including offline support, installability on mobile devices, and push notifications. GenMB can automatically convert generated apps into PWAs by creating a manifest.json, service worker, and offline fallback page.
An AI pattern that improves response quality by first retrieving relevant documents from a knowledge base, then passing them as context to the language model. GenMB's Vector DB service enables RAG in generated apps: upload documents, auto-chunk them, generate embeddings, and query semantically. Used in GenMB Assistants for knowledge base Q&A.
A security model that restricts system access based on user roles (admin, editor, viewer) rather than individual permissions. Each role defines what actions a user can perform. GenMB provides a built-in RBAC service with a client SDK that generated apps can use to manage roles, check permissions, and gate features.
A web design approach where layouts adapt to different screen sizes — desktop, tablet, and mobile — using CSS media queries, flexible grids, and relative units. GenMB generates responsive applications by default, using Tailwind CSS utility classes that handle breakpoints automatically.
An API architecture style that uses standard HTTP methods (GET, POST, PUT, DELETE) to perform CRUD operations on resources identified by URLs. REST is the most common API pattern for web applications. GenMB generates RESTful backend endpoints for full-stack applications and uses REST internally for its service communication.
Software delivered over the internet on a subscription basis, rather than installed locally. Users access SaaS products through a browser (Gmail, Slack, Notion). GenMB is itself a SaaS product and can generate SaaS-style applications with user authentication, subscription logic, dashboards, and admin panels.
A formal definition of data structure — what fields exist, their types, constraints, and relationships. Database schemas define tables and columns. API schemas define request and response shapes. GenMB's Schema Designer lets users visually define data models that the AI uses to generate database tables, CRUD operations, and type-safe code.
A package of tools, libraries, and documentation that simplifies building on a specific platform or service. GenMB provides 9 client-side SDKs (file storage, auth, chatbot, contact form, data client, RBAC, vector DB, assistant widget) that are automatically injected into generated apps when the corresponding service is detected.
GenMB's built-in OWASP Top 10 static analysis tool that runs automatically during code generation. It detects cross-site scripting (XSS), SQL injection, exposed secrets, misconfigured permissions, insecure authentication, and insecure communications. Findings are automatically remediated by Code Healer.
A cloud computing model where the provider manages all infrastructure and automatically scales compute resources. Developers deploy functions or containers without provisioning servers. "Serverless" doesn't mean no servers — it means you don't manage them. GenMB deploys apps to Cloud Run, a serverless container platform that scales to zero when idle.
A rendering strategy where HTML pages are generated at build time rather than on each request. The resulting static files are served from a CDN, making SSG sites extremely fast and cheap to host. Best suited for content that changes infrequently — blogs, documentation, marketing pages.
An authentication method that lets users log in to multiple applications using one set of credentials. GenMB supports OIDC-based SSO with JIT (Just-in-Time) provisioning for Enterprise customers. Once configured for your email domain, all employees are automatically authenticated and provisioned into your GenMB workspace without individual account setup.
A rendering strategy where the server generates the full HTML for each page request before sending it to the browser. SSR improves initial load performance and SEO because search engines receive complete HTML instead of an empty shell that requires JavaScript to populate. Next.js (used by GenMB's frontend) supports SSR natively.
The practice of managing and synchronizing data (state) across different parts of a user interface. As applications grow, tracking what data lives where becomes complex. Common approaches include React Context, Redux, and Zustand. GenMB's frontend uses React hooks and Context for state management without external state libraries.
A prefix added before a domain name (e.g., yourapp.genmb.com). GenMB offers free subdomain deployment where every app gets a unique subdomain at *.genmb.com with automatic HTTPS. Available on all plans including Free. Custom domains (myapp.com) are available on Pro and Business plans.
Deploying an application to a subdomain of the platform's domain (e.g., yourapp.genmb.com). GenMB provides one-click subdomain deployment with automatic SSL/HTTPS, available on all plans including the free tier.
A specialized database that stores data as high-dimensional numerical vectors (embeddings) instead of rows and columns. Vector databases enable semantic similarity search — finding content by meaning rather than exact keywords. GenMB's Vector DB service uses pgvector (a PostgreSQL extension) with automatic text chunking and embedding generation, making it easy to add semantic search or RAG to generated apps.
A system for tracking changes to code over time. GenMB provides built-in version history with automatic version creation on every change, side-by-side comparison, and atomic rollback to any previous state. For external version control, GenMB also offers GitHub Sync for pushing and pulling code.
A development approach where you describe what you want to build in natural language and let AI generate the code. Coined by Andrej Karpathy, vibe coding shifts the focus from syntax and implementation details to intent and desired outcomes. GenMB is a purpose-built platform for vibe coding, generating complete full-stack applications from text prompts.
A GenMB feature that lets users click elements in the live preview to modify them visually. Users can change text, colors, spacing, sizing, and layout without touching the code editor. All changes are reflected in the source code automatically.
A fast JavaScript build tool that bundles and transpiles modern code (TypeScript, JSX) for production. GenMB uses Vite automatically for React + TypeScript apps, since TSX cannot run in a browser without transpilation. Vite handles tsconfig generation, TypeScript compilation, tree-shaking, and creates an optimized production bundle. Vanilla JS and React apps don't require a build step.
An HTTP callback that automatically sends data to a specified URL when an event occurs. Unlike APIs where you poll for updates, webhooks push data to you in real time. For example, Stripe sends a webhook when a payment succeeds. GenMB's workflow system can trigger actions on webhook events.
A communication protocol that enables persistent, bidirectional data exchange between a browser and server over a single connection. Unlike HTTP (request-response), WebSockets allow the server to push updates instantly. GenMB uses WebSockets for real-time code generation streaming, agent mode progress, chat refinement, and terminal sessions.
A visual automation pipeline in GenMB that connects triggers to actions using a node-based editor. Workflows support 16+ node types including AI processing, HTTP requests, conditional logic, loops, email, Slack, database queries, and code execution. Workflows can be triggered manually, on a schedule (cron), or via webhook. Validated as a DAG (directed acyclic graph) to prevent circular dependencies. Available on Business plan.
Put these concepts into practice. Describe your app idea and let GenMB generate the code.
Try GenMB Free