Frameworks, apps, and products where AI dynamically generates UI at runtime.
This list focuses on projects you can study, run, or use, with an emphasis on:
- Open-source frameworks and SDKs.
- Reference implementations and example apps.
- Public products that clearly ship generative UI patterns.
It intentionally avoids becoming a generic AI tools directory.
- What is Generative UI
- Inclusion Criteria
- Libraries and Frameworks
- Reference Apps and Open-Source Products
- Research and Experimental Projects
- Public Working Models
Generative UI is the idea that an AI system does not only generate text or code, but also dynamically decides what interface to render for a given task, user intent, or context.
In practice, that can mean:
- A model choosing UI components and their props at runtime.
- An agent returning structured UI instead of only markdown text.
- A system generating interactive layouts, forms, tools, or visualizations tailored to the prompt.
Projects on this list should meet at least one of these:
- They are a notable open-source repository directly related to generative UI.
- They are a strong implementation reference with runnable code.
- They are a public product or demo that clearly exposes generative UI behavior.
Projects are excluded if they are:
- Generic AI app frameworks with no real UI-generation angle.
- Low-signal toy repos without meaningful implementation value.
- Unrelated infrastructure, model hosting, vector databases, or agent tooling.
- CopilotKit - Framework for building agent-native applications with generative UI, shared state, and human-in-the-loop workflows.
- Vercel AI SDK - General AI app toolkit and one of the core building blocks behind many generative UI implementations.
- AG-UI Protocol - Open protocol for bringing agents into frontend applications, including structured messages and generative UI primitives.
- tambo - React SDK focused explicitly on generative UI, where the AI decides which components to render and how to configure them.
- assistant-ui - React library for AI chat with structured outputs, tool rendering, approvals, and composable AI interfaces.
- OpenUI - Open standard for generative UI centered on agentic interface rendering.
- Hashbrown - Framework for building agents that run in the browser, with support for Angular and React.
- react-native-gen-ui - React Native library for function-calling-driven UI rendering and streaming chat experiences on mobile.
- mdocui - Markdown-plus-tag streaming UI library for LLM output, aimed at mixing prose and interactive components in one response stream.
- Morphic - AI-powered search engine with generative UI and a strong open-source implementation for adaptive search experiences.
- Cofounder - Full-stack app generator with a generative UI angle, covering interface generation alongside broader product scaffolding.
- gemini-chatbot - Vercel Labs template showing dynamic chat and generative UI patterns using Gemini and the AI SDK.
- gen-ui-python - Python-backed generative UI example useful for teams building with LangChain Python.
- gen-ui - LangChain.js and Next.js template for building a generative UI application with prebuilt UI features.
- Google Research: Generative UI - Research project and paper site describing a full-stack implementation where the model generates custom web pages for prompts.
- Google Search AI Mode - Search uses Gemini to generate dynamic visual layouts, interactive tools, and simulations tailored to the query.
- Gemini app generative interfaces - First generative interfaces inside the Gemini app, including visual layouts and dynamic views.
- Bolt.new - StackBlitz product for prompt-to-app generation with live preview and rapid iteration in the browser.
- Lovable - Prompt-to-app builder focused on turning product ideas into deployable web applications.
- Figma Make - Figma's prompt-to-interface product with generative UI as its core interaction model.
- v0 - Vercel's production-oriented prompt-to-app product and one of the clearest public examples of generative UI as a product category.
Please read CONTRIBUTING.md before opening a pull request.
Good additions are:
- Clearly relevant to generative UI.
- Open-source or publicly usable.
- Technically informative, not just trend-chasing.
- Concise enough to describe in one line.
- This list is independent and unofficial.
- Similar naming may exist elsewhere on the web; this repository is intended as a GitHub-first curated list.
- If a project is adjacent but not central to generative UI, it should stay out until the connection is clear.