NLUX

The Powerful Conversational AI JavaScript Library ✨💬

Docs Website | Discord Community | X

Do you like this project ? Please star the repo to show your support 🌟 🧡
Building with NLUX ? Get in touch — We'd love to hear from you.

---

NLUX (for Natural Language User Experience) is an open-source JavaScript and React JS library that makes it super simple to integrate powerful large language models (LLMs) like ChatGPT into your web app or website. With just a few lines of code, you can add conversational AI capabilities and interact with your favourite LLM.

Key Features 🌟

• Build AI Chat Interfaces In Minutes ― High quality conversational AI interfaces with just a few lines of code.
• React Components & Hooks ― <AiChat /> for UI and useChatAdapter hook for easy integration.
• LLM Adapters ― For ChatGPT ― LangChain 🦜 LangServe APIs ― Hugging Face 🤗 Inference.
• A flexible interface to Create Your Own Adapter 🎯 for any LLM ― with support for stream or batch modes.
• Assistant and User Personas ― Customize the assistant and user personas with names, images, and descriptions.
• Streaming LLM Output ― Stream the chat response to the UI as it's being generated.
• Custom Renderers ― Render AI messages with custom components inside the chat interface.
• Highly Customizable ― Tune almost every UI aspect through theming, layout options, and more.
• Zero Dependencies ― Lightweight codebase ― Core with zero dependency and no external UI libraries.

Repo Content 📦

This GitHub repository contains the source code for the NLUX library.
It is a monorepo that contains code for following NPM packages:

⚛️ React JS Packages:

• @nlux/react ― React JS components for NLUX.
• @nlux/langchain-react ― React hooks and adapter for APIs created using LangChain's LangServe library.
• @nlux/openai-react ― React hooks for the OpenAI API, for testing and development.
• @nlux/hf-react ― React hooks and pre-processors for the Hugging Face Inference API
• @nlux/nlbridge-react ― Integration with nlbridge, the Express.js LLM middleware by the NLUX team.

🟨 Vanilla JS Packages:

• @nlux/core ― The core Vanilla JS library to use with any web framework.
• @nlux/langchain ― Adapter for APIs created using LangChain's LangServe library.
• @nlux/openai ― Adapter for the OpenAI API, for testing and development.
• @nlux/hf ― Adapter and pre-processors for the Hugging Face Inference API.
• @nlux/nlbridge ― Integration with nlbridge, the Express.js LLM middleware by the NLUX team.

🎁 Themes & Extensions:

• @nlux/themes ― The default Luna theme and CSS styles.
• @nlux/markdown ― Markdown stream parser to render markdown as it's being generated.
• @nlux/highlighter ― Syntax highlighter based on Highlight.js.

Please visit each package's NPM page for information on how to use it.

Docs & Examples 🤩

For stable v1.x — Developer documentation, examples, and API references are available at:
nlux.dev

For v2.x-beta — The docs website is being built and will be available at:
docs.nlkit.com

For v2.x-beta, please refer to TS type definitions, specs, and source code if as the documentation is incomplete and still being written.

Design Principles ⚜️

The following design principles guide the development of NLUX:

• Intuitive ― Interactions enabled by NLUX should be intuitive. Usage should unfold naturally without obstacles or friction. No teaching or thinking should be required to use UI built with NLUX.

• Performance ― NLUX should be as fast as possible. Fast to load, fast to render and update, fast to respond to user input. To achieve that, we should avoid unnecessary work, optimize for performance, minimize bundle size, and not depend on external libraries.

• Accessibility ― UI built with NLUX should be accessible to everyone. It should be usable by people with disabilities, on various devices, in various environments, and using various input methods (keyboard, touch, voice).

• DX ― NLUX recognizes developers as first-class citizens. The library should enable an optimal DX (developer experience). It should be effortless to use, easy to understand, and simple to extend. Stellar documentation should be provided. The feature roadmap should evolve aligning to developer needs voiced.

Mission 👨‍🚀

Our mission is to enable developers to build outstanding LLM front-ends and applications, cross platforms, with a focus on performance and usability.

Community & Support 🙏

• Star The Repo 🌟 ― If you like NLUX, please star the repo to show your support.
  Your support is what keeps this open-source project going 🧡
• GitHub Discussions ― Ask questions, report issues, and share your ideas with the community.
• Discord Community ― Join our Discord server to chat with the community and get support.
• nlux.dev Developer Website ― Examples, learning resources, and API reference.

License 📃

NLUX is licensed under Mozilla Public License Version 2.0 with restriction to use as part of a training dataset to develop or improve AI models, or as an input for code translation tools.

Paragraphs (3.6) and (3.7) were added to the original MPL 2.0 license.
The full license text can be found in the LICENSE file.

In a nutshell:

• You can use NLUX in your personal projects.
• You can use NLUX in your commercial projects.
• You can modify NLUX and publish your changes under the same license.
• You cannot use NLUX's source code as dataset to train AI models, nor with code translation tools.

Wondering what it means to use software licensed under MPL 2.0? Learn more on MPL 2.0 FAQ.
Please read the full license text in the LICENSE file for details.

About The Developer 👨‍💻

NLUX is a new open-source project that's being led by Salmen Hichri, a senior front-end engineer with over a decade of experience building user interfaces and developer tools at companies like Amazon and Goldman Sachs, and contributions to open-source projects.