Key Features
Tech Stack
> The LLM generates documentation
so, not documentation? Why not write your own engine and detect the official docs? e.g. docs.rs would do this wonderfully
Looking up official documentation would require shipping sophisticated parsers for each language, plus a way to map tokens to their corresponding docs. I'd also need to maintain those mappings as libraries evolve and documentation moves. Some ecosystems make this easier (Rust with docs.rs), but others would be much harder (AWS documentation, for example).
I also want explanations visible directly in hover hints rather than linking out to external docs. So even with official documentation, I'd need to extract and present the relevant content anyway.
Beyond that, the LLM approach adapts to context in ways static docs can't. It generates explanations for code within the documentation you're reading, and it works on code that doesn't compile, like snippets with stubs or incomplete examples.
It could be interesting in the future to look into doing some type of hybrid approach where an LLM goes out and searches up the documentation, that way it's a little bit more flexible. But that would also be a bit slower and more costly.
For whom? The whole reason I want to consult docs is to get the official documentation on a given topic. How could I trust anything it says, and what’s to say any earned trust is durable over time?
what is the name for this kind of pointless, lazy, selective, quoting that willfully misconstrues what's being quoted? the answer to this question is incredibly clear: for the developer that created this tool. if that makes you unhappy enough to malign them then maybe you should just not use it?
That said, people have built this without LLMs years, even decades, ago. But UX has fallen by the wayside for quite some time in the companies that used to build IDEs. Then some fresher devs come along and begin a project without the benefit of experience in a codebase with a given feature … and after some time someone writes a plugin for VSCode to provide documentation tooltips generated by LLM because “there is just no other way it can be done.”
We have language servers for most programming languages. Those language servers provide the tokens one needs to use when referencing the documentation. And it would be so much faster than waiting for an LLM to get back to you.
TBH, if anyone’s excuse is “an LLM is the only way to implement feature Q,” then they’re definitely in need of some experience in software creation.
In my opinion the shipped product is better than the unshipped product. While of course I would prefer the version that you have designed, I sure don't have time to build it, and I'm guessing you don't either.
If this was our day jobs and we were being paid for it, it would be a much different story, but this is a hobby project made open source for the world.
Some things that might be useful to know to speed you up:
1. Most code blocks on the internet are easy to find in a webpage. They generally are surrounded by `<code/>` tags. You can query for these using the method in my extension. Then you will need to filter out any code block that has a `<span/>` count <= 1, as code blocks are used to highlight arbitrary stuff on the internet and you would really just want to find the blocks of code.
2. You will need a method to identify when to generate documentation for a code block even with your implementation, as some documentation websites are one really long page. You can do this with two types of observers that I use in my codebase. One will identify when a code block is in view and the other will keep track of mutations to code blocks. You need to keep track of mutations because sites like ChatGPT continuously edit a code block while streaming a response. You want to generate documentation once it's done, as that's when the code is well formed. I have a janky example of how to do this in my extension. Claude should be able to find the code.
3. LLMs were useful for building this, but they struggled with design decisions, especially around UX. This project seems out of distribution for them. Claude probably won't suggest the right solution a lot of the time, but if you have it list out multiple options, it can usually identify which one is best.
I will probably check back here in a couple weeks if I don't hear anything from you. I would be really impressed if you can get something working in a weekend that is *not just hard coded for a specific use case, but scales well.*
To clarify my earlier point, I wasn't suggesting this is impossible, just that it's not *practical* to build a universal LSP that works with every language and framework out of the box without anything local to index. I don't think an reusing an LSP would be a great fit here either, since LSPs rely on having full project context, dependencies, and type information. These aren't available when analyzing code snippets on arbitrary webpages.
Parsing was never my major concern though. It's the "map tokens to URLs" part. A universal mapping for every token to every piece of documentation on the internet is *impractical* and difficult to maintain. To achieve parity without LLMs, I'd need to write and maintain parsers for every documentation website, and that assumes documentation even exists for most tokens (which it doesn't).
I think kristopolous's suggestion of grounding the LLM with data sources that keep a serialized database of documentation from many different places makes the most sense. That way, the LLM is just extracting and presenting key information from real documentation rather than generating from scratch.
There are probably ways to make this easier. Maybe an offline job that uses LLMs to keep mappings up to date. The project could also be scoped down to a single ecosystem like Rust where documentation is centralized, though that falls apart once you try to scale beyond one language as mentioned above. Something similar could work here: use the raw definition on GitHub combined with an LSP to generate information.
Open to other suggestions on how to bridge this gap.
They quoted the part they were replying to. The point was to show what they were asking about. If your question pertains to only a part of the text, it only makes sense to be selective. That's not wilfully misconstruing anything; that’s communicating in a clear, easy-to-follow way. The context is still right up there for reading, for anyone who needs to review it.
> the answer to this question is incredibly clear: for the developer that created this tool
Questions aren't only ever asked out of pure curiosity; sometimes they're asked to make the other person give them more consideration. The question you quote was accompanied by an explanation of how the commenter found the approach less simple for them as a user, suggesting that perhaps they think the developer would have done better to consider that a higher priority. (I might add that you, too, chose to selectively omit this context from your quoting—which I personally don't see as problematic on its own, but the context does require consideration, too.)
> if that makes you unhappy enough to malign them then maybe you should just not use it?
The author of the extension chose to share what they made for others to use. They asked for feedback on user experience and expressed doubt about their design decisions. If someone finds they might not want to use it because of what they consider fundamentally flawed design, why couldn’t they tell the author? It's not like they were rude or accused them of any wrong-doing (other than possibly making poor design choices).
You could just token match (use tree-sitter or something similar) and fetch the official docs. Keep it dead-simple so there's no way you can provide false positives (unlike what's happening rn where hallucinations will creep in).
> It generates explanation
Again, I don't want that. It's not a feature, it's a limitation that right now gives you fake information.
I could see getting actual docs being useful. Spitting out the delusions of an LLM is pretty well covered already, at least in my stack.
Also look into https://cht.sh/
Remember: incorrect (misleading) documentation is worse than no documentation.
What this might be better for is use-cases that don't require extreme precision. Imagine it for learning language or reading sophisticated academic literature. For example, https://archive.org/details/pdfy-TJ7HxrAly-MtUP4B/page/n111/...
Stuff like that is hard and every tool to make the complicated more legible I'd embrace.
curl cht.sh
it's a wiki system.
Not affiliated with Hacker News or Y Combinator. We simply enrich the public API with analytics.