Improve readme wording

Author: immitsu

readme.md

@@ -2,9 +2,9 @@
 
 A simple parser for the Netscape Bookmark File Format (NBFF), commonly generated by browsers when exporting bookmarks.
 
-It supports parsing bookmarks into various formats — including customizable structures — and converting the parsed data back into an HTML string.
+It supports parsing into various formats — including customizable structures — and writing the data back.
 
-Additionally, the parser can merge the contents of multiple bookmark files into a single, unified string.
+It also supports merging multiple files.
 
 ## Features
 
@@ -22,7 +22,6 @@ Additionally, the parser can merge the contents of multiple bookmark files into
   - [stringify](#stringify)
   - [flatStringify](#flatstringify)
   - [merge](#merge)
-- [Attribute Handling](#attribute-handling)
 - [CLI](#cli)
   - [exclude](#exclude)
   - [merge](#merge-1)
@@ -45,6 +44,8 @@ yarn add nbff-parser
 
 The parser expects HTML file content to be provided as a string.
 
+Returns bookmark attributes with lowercase names and normalized values.
+
 ### `parse`
 
 [Type definition](./types/parse/parse.d.ts)
@@ -85,10 +86,10 @@ const bookmarks = parse(html)
 
 You can configure the output by activating options passed as the second argument.
 
-| Option         | Type       | Description                                                                                 |
-| -------------- | ---------- | ------------------------------------------------------------------------------------------- |
-| `excludeAttrs` | `string[]` | Excludes specified attributes from output. See [attributes definition](./types/attrs.d.ts). |
-| `withId`       | `boolean`  | Adds hierarchical identifiers `id` and `pid` to each item.                                  |
+| Option         | Type            | Description                                                                      |
+| -------------- | --------------- | -------------------------------------------------------------------------------- |
+| `excludeAttrs` | `AllAttrKeys[]` | Excludes specified attributes from output. See [definition](./types/attrs.d.ts). |
+| `withId`       | `boolean`       | Adds hierarchical identifiers `id` and `pid` to each item.                       |
 
 ### `flatParse`
 
@@ -135,11 +136,11 @@ const bookmarks = flatParse(html)
 
 You can configure the output by activating options passed as the second argument.
 
-| Option         | Type                        | Description                                |
-| -------------- | --------------------------- | ------------------------------------------ |
-| `excludeAttrs` | `string[]`                  | Excludes specified attributes from output. |
-| `withId`       | `boolean`                   | Adds incremental numeric `id` to items.    |
-| `transform`    | `(item: FlatBookmark) => T` | Transforms items, omitting falsy returns.  |
+| Option         | Type                        | Description                                                                      |
+| -------------- | --------------------------- | -------------------------------------------------------------------------------- |
+| `excludeAttrs` | `AllAttrKeys[]`             | Excludes specified attributes from output. See [definition](./types/attrs.d.ts). |
+| `withId`       | `boolean`                   | Adds incremental numeric `id` to items.                                          |
+| `transform`    | `(item: FlatBookmark) => T` | Transforms items, omitting falsy returns.                                        |
 
 ### `customParse`
 
@@ -155,10 +156,10 @@ The methods described above rely on it internally.
 import { customParse } from 'nbff-parser'
 
 const handlers = {
-  addBookmark,
-  describeBookmark,
-  openFolder,
-  closeFolder
+  openFolder, // <H1>, <H3>
+  addBookmark, // <A>
+  describeBookmark, // <DD>
+  closeFolder // </DL>
 }
 
 const bookmarks = customParse(html, handlers)
@@ -184,7 +185,7 @@ const stringified = stringify(parsed)
 
 Converts the flat list from `flatParse` back into an HTML string.
 
-> It requires using `flatParse` with `{ withId: true }` to ensure unique item IDs, and without `transform` to ensure a known structure.
+> It requires using `flatParse` with `{ withId: true }` to ensure unique item IDs.
 
 ```js
 import { flatParse, flatStringify } from 'nbff-parser'
@@ -198,16 +199,18 @@ const stringified = flatStringify(parsed)
 
 [Type definition](./types/merge/merge.d.ts)
 
-Merges the contents of multiple HTML files into a single HTML string.
+Merges parsed files into a single HTML string.
 
-## Attribute Handling
+```js
+import { merge } from 'nbff-parser'
 
-- Attribute names are returned lowercased.
-- Attribute values may be slightly normalized.
-- See detailed attribute types [here](./types/attrs.d.ts).
+const merged = merge(html1, html2, ...)
+```
 
 ## CLI
 
+CLI methods work with actual files.
+
 Usage:
 
 ```sh
@@ -224,8 +227,6 @@ Removes specified attributes from the HTML file.
 | `attrs=attr1,attr2,...` | Comma-separated list of attributes to exclude        | Required |
 | `output=path/to/output` | Path to save the output file; defaults to input file | Optional |
 
-If `output` is not provided, the changes will overwrite the original file.
-
 Example:
 
 ```sh