An Open-Source Path to ChatGPT-Style Interactive Answers: OpenGenerativeUI
OpenGenerativeUI is an open-source path to ChatGPT-style interactive answer UIs in your own product — verified identity, features, official repo, and docs.
Verified account Atai Barkai (@ataiiam) posted on October 8, 2026 that ChatGPT had rolled out Intelligent UI, and introduced OpenGenerativeUI as the open-source path to the same interactive-answer capability in your own product. A follow-up post linked the official repository directly, and the repo description plus official docs confirm the feature scope and rendering approach.

Image source: CopilotKit / GitHub
OpenGenerativeUI is a CopilotKit-based open-source showcase: an agent generates live HTML/SVG, and the frontend renders it inside a sandboxed iframe. Search evidence describes a setup working with LangChain Deep Agents, while related docs describe a LangGraph-based configuration in the same family. The core idea is simple: instead of text-only replies, the agent builds fully interactive visual components right inside the chat.
What it builds and who it serves
The repo description lists algorithm visualizations, 3D animations, charts and diagrams, and interactive widgets. The original post directly listed algorithm visualizations, charts and diagrams, and interactive widgets; 3D animations appear in the repo description, not in the original post.
- Algorithm visualizations: binary search, BFS vs DFS, sorting algorithms
- 3D animations: interactive WebGL/CSS3D scenes
- Charts & diagrams: pie charts, bar charts, network diagrams
- Interactive widgets: forms, simulations, math plots
It fits products that need dashboards or visual explanations generated directly inside a conversation, such as learning tools or data-heavy assistants. One user in the thread said it saved significant time for dashboards, but that is a personal report — results will vary by setup.
Verified usage path and docs
The full install command was not included in this verification bundle, so follow the official repo README for installation. The canonical source is https://github.com/CopilotKit/OpenGenerativeUI.
The official docs page (https://docs.copilotkit.ai/generative-ui/open-generative-ui) explains the runtime behavior: enabling openGenerativeUI in the runtime streams agent-authored HTML and CSS to the built-in OpenGenerativeUIActivityRenderer, which mounts it inside a sandboxed iframe. The docs present a minimal setup that starts with just chat — no custom sandbox functions or custom tools.
Why it matters now, and verified limits
The ChatGPT-side announcement is cross-confirmed by the official @ChatGPT post: Intelligent UI makes everyday answers more visual, complex topics easier to grasp, and task-solving tools available on the spot. That is OpenAI's marketing framing, so read it as an attributed claim.
On the OpenGenerativeUI side, the author's "months ago" open-sourcing remark is a relative claim — without a stated first-release date in the bundle, the exact number of months cannot be pinned down. The repo issue tracker also shows ongoing work such as visualization quality regressions and template rework, so check open issues before adopting.
Sources
- CopilotKit/OpenGenerativeUI official repo: https://github.com/CopilotKit/OpenGenerativeUI
- CopilotKit official docs (Open Generative UI): https://docs.copilotkit.ai/generative-ui/open-generative-ui
- Atai Barkai (@ataiiam) original post: https://x.com/ataiiam/status/2107933770926195074
- ChatGPT (@ChatGPT) Intelligent UI announcement: https://x.com/ChatGPT/status/2107933133635248145