---
title: "Enhancements"
---

> Documentation Index
> Fetch the complete documentation index at: https://kmworks.date/llms.txt
> Use this file to discover all available pages before exploring further.

# Enhancements

Improvements KMServer adds on top of Java parity, plus intentional behavior differences. Accepted gaps that are not planned to be closed are listed under [Known limitations](/server/limitations).

## Capabilities the Java version does not have

- [Outbound webhooks](/server/enhancements/webhooks): library events POSTed as JSON to configured URLs, with HMAC-SHA256 signatures
- [History events](/server/enhancements/history-events): scan-triggered trashing and trash emptying recorded in `GET /api/v1/history`
- [Search](/server/search): simplified ↔ traditional Chinese cross-search and CJK boundary unigrams
- [EPUB 2 series metadata](/server/enhancements/epub2-metadata): calibre-style OPF 2 `<meta>` fallback for series title and number
- [Natural sort](/server/enhancements/natural-sort): numbered titles sort by numeric value, with a configurable ICU locale
- [komf integration](/server/enhancements/komf): one-click setup of a komf metadata fetcher
- [Thumbnail file storage](/server/enhancements/thumbnail-storage): opt-in file-backed thumbnails instead of DB blobs
- [komga-riir databases](/server/enhancements/komga-riir): adopt databases created by komga-riir

## Behavior differences

Intentional differences from the Java behavior:

- Book and series thumbnail endpoints answer `If-None-Match` against a deep ETag computed from the thumbnail row (blob rows: row id + file size; file-backed rows: file mtime + size), so a cache revalidation returns 304 without hashing the response body; file-backed rows (sidecars, `thumbnails.storage = "file"`) are revalidated with a file stat instead of a full read. The Java version always hashes the full response body (shallow ETag); `thumbnails.deep-etag = false` (env `KOMGA_THUMBNAILS_DEEPETAG`) switches back to that behavior. ETag values are opaque and not interchangeable between the two implementations — switching either way just costs one fresh 200 per image.
- Page order and series book numbering are case-sensitive; the Java natural comparator is case-insensitive.
- `GET /actuator/sessions` without `?username=` lists all sessions; the Java version rejects the request, as `username` is a required parameter there.
- `/api/v1/books/{bookId}/next` and `/previous` break `numberSort` ties by book id, matching the On Deck selection order; the Java version orders and seeks by `numberSort` alone, so navigation between books with duplicate `numberSort` is unstable and can diverge from On Deck ([gotson/komga#2182](https://github.com/gotson/komga/pull/2182)).

Source: https://kmworks.date/server/enhancements/index.mdx