Replies: 9 comments 1 reply
-
|
Would love to see a complete abstraction for streaming. It has become the default way of interacting with llms for anything with a longer response time. I think we've got some good building blocks in Angular but we need something that hides all the complexity and different streaming strategies. The resource API is capable of handling this, but there's still too much that's required of authors to consume the stream. |
Beta Was this translation helpful? Give feedback.
-
|
@markgoho I agree - which is why @MikeRyanDev and I built a TS framework that uses the resource API so you can use LLMs in your Angular apps. Shameless plug I know, but I thought I would share! Check it out: https://github.com/liveloveapp/hashbrown |
Beta Was this translation helpful? Give feedback.
-
|
Rich Chat Input Control Currently, there is no single Angular library that provides a complete solution for: Trigger-based dropdowns (e.g., @, #, /) for mentions, commands, or tags. In contrast, React and other competitive frameworks offer multiple mature libraries (like react-mentions, draft-js, tiptap, etc.) that cover these use cases out of the box. Angular developers are left to either stitch together multiple packages or build custom solutions from scratch. 🧩 Proposed FeaturesTrigger-based Suggestions
Markdown Support
Customizability
Performance & Accessibility
🛠️ Personal ExperienceI recently needed this functionality for a chat interface and couldn’t find a single Angular package that met the requirements. I ended up building my own library, but faced several challenges around performance, extensibility, and markdown rendering. A well-designed, CDK-compliant solution would be a huge win for the Angular ecosys 7440 tem. |
Beta Was this translation helpful? Give feedback.
-
|
We're working on generative ui, and we've built a basic solution using createComponent recursively, but it would be super cool if there was some kind of established pattern/solution for describing UIs using data. On the business side, not only is it useful for LLM's generating UIs, but it also enables low-code/no-code solutions where portions of an app can be built without source code and saved to a CMS. |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
-
|
We could consider an infinite reverse scroll for the message listing when the chat becomes large as well as elements for streaming typing that support markdown. |
Beta Was this translation helpful? Give feedback.
-
|
In my case, I’m working on an educational platform (GradeBase) where we use Angular on the frontend. A potential AI integration could be an assistant that suggests usability improvements based on students’ grades and progress. This could be implemented with an LLM connected through Angular’s CDK, enabling contextual recommendations directly in the UI. |
Beta Was this translation helpful? Give feedback.
-
|
Answer 1: Currently building a search feature. Have a customer support feature on our roadmap. Answer 2: Saving time. Reduced interaction with junior staff members who can't address customer queries swiftly and with precision. Senior staff members are better than AI but there's not enough of them to go around. Answer 3: Existing LLMs mix different Angular versions up. When they can't find a concrete solution to a problem they start lying, combining together parts of different solutions and presenting them as the most optimal one in the entire world, while it's actually a bunch of nonsense - keywords, interfaces, decorators etc. that do not exist in the API. Answer 4: We would love to have AI as part of Angular Services (e.g. the add-on for Visual Studio Code). As for the features - let's start with the basics and expand from there - generate code snippets for common tasks (e.g. HTTP interceptor, Web Worker implementation, basic data store structure etc.), assist with code analysis (e.g. point out possible areas of improvement but not the simple stuff, and rather ways to optimize Angular code so that it's performant and using the latest syntax). Guides on how to integrate AI into an existing project are a must. Any and all tools must come from the Angular team and not third parties. |
Beta Was this translation helpful? Give feedback.
-
|
Not directly related to the given discussion questions, but I feel this is related enough:
For me, the testing situation is frustrating. It makes me question, whether to choose Angular for professional greenfield projects. I feel, there is currently no way to write low level tests, whose framework is not already outdated at the time of initial writing: Karma deprecated but still shifts in default scaffold schema, Jest experimental, Vitest experimental - what should I use? (By the way, thanks for all your great work! I really love the direction Angular is going since its v17 renaissance. All your reaching out to the community and collecting feedback. Doing such huge changes to Angular in always easy migrate-able ways to prior versions. That is great work I am very grateful for 🙏❤️🙏) |
Beta Was this translation helpful? Give feedback.
-
|
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
Uh oh!
There was an error while loading. Please reload this page.
-
Authors:
Gemini@alxhub & @pkozlowski-opensourceArea: Angular Framework
Posted: September 12, 2025
Status: Open
AI is shifting how web applications are built, but also the nature of applications themselves. User experiences are rapidly evolving as developers incorporate generative AI directly in their UIs. Things that were once considered futuristic, such as interactive assistants, natural language query systems, generative user interfaces, and interactive tutorials are within the realm of possibility.
We want to understand how our community is approaching this shift and where we can provide meaningful support through core primitives, tooling, or helpers in our Component Development Kit (
@angular/cdk).The goal of this RFC is not to propose a specific API, but rather to solicit broad feedback that will help shape our roadmap and future-proof Angular.
Discussion Question 1: What AI-driven features are you currently building, planning for your roadmap, or dreaming of building inside your Angular applications?
Discussion Question 2: For the features in (1), what is the core user need or business problem they solve? Why is integrating the AI feature directly into the client application important for the user experience?
Discussion Question 3: What technical challenges or friction have you experienced while integrating AI/LLM SDKs and services into your Angular applications? Which libraries or utilities have you tried to use? What was the experience like?
Discussion Question 4: Looking forward, how can the Angular framework best support your development of these features? What features, APIs, tools, or guides would most reduce friction and accelerate your roadmaps?
Beta Was this translation helpful? Give feedback.
All reactions