hast-util-to-string

This package is a perfect example of a focused utility that solves one problem exceptionally well. The API is dead simple: import `toString()` and pass it a hast node. It extracts plain text from HTML AST structures, handling all the edge cases you'd expect (script/style tags excluded, proper whitespace handling). TypeScript support is excellent with proper type definitions that work seamlessly with the unified/rehype ecosystem.

The documentation is clear and complete despite being concise. Examples cover the common cases, and the behavior is predictable enough that you rarely need to reference docs after first use. Error handling is straightforward - it won't crash on malformed input, just returns what it can extract. Integration with rehype plugins is seamless since it uses standard hast types.

In daily use, it's one of those utilities you import, use once, and forget about because it just works. Whether you're building markdown processors, content extractors, or SEO tools that need plain text from HTML ASTs, this handles the grunt work reliably. The small API surface means there's nothing to learn beyond the single function call.

This package does exactly one thing: extract plain text from hast (HTML Abstract Syntax Tree) nodes. The API is ridiculously simple - just import the function and pass it a node. I've used it in several projects processing markdown-to-HTML pipelines and it's never surprised me. The function handles edge cases intelligently, like stripping out script and style tag content by default, which is exactly what you want 99% of the time.

The documentation is minimal but that's because there's genuinely not much to explain. The README has enough examples to understand the whole API in 30 seconds. Error messages are straightforward when you pass invalid input. It integrates seamlessly with the unified/remark/rehype ecosystem, which is clearly its intended use case.

Debugging is straightforward because the package is pure utility with no hidden state. When something seems wrong, it's usually your hast tree structure, not this package. The TypeScript types are solid and helpful. For such a focused utility, it's hard to imagine how it could be better.

This package is the epitome of a well-designed single-purpose utility. You pass it a hast node (HTML Abstract Syntax Tree), and it returns the plain text content. No configuration needed, no surprises. I've used it across multiple projects dealing with unified/remark/rehype ecosystems, and it just works every time.

The learning curve is essentially zero—if you're already working with hast trees, you literally import and call it. The documentation is clear and includes practical examples showing edge cases like handling whitespace. Error messages are straightforward JavaScript errors when you pass invalid input, which makes sense given the simplicity.

What I appreciate most is how it handles the nitty-gritty details correctly: script/style tags are ignored, whitespace is normalized properly, and it recursively processes nested structures without me having to think about it. When debugging AST transformations, this is invaluable for quickly checking what text content you're actually working with. It's one of those packages that becomes invisible in your workflow because it never causes problems.