Nice write-up about yourei.jp. Yeah, I tended to use it more than weblio.jp as well.

For ficton-based sentences, I also like using massif.la since its pulling sentences from web novels on syosetsu.com. The only con since many of the stories are written by aspiring writers (ie. amateurs), there's no guarantee everything is proofread and correct, but most of the time, it's been fine.

I also find it helpful that these writers may overly rely on cliches and phrasing, as it's better from a learner standpoint to see how a word is most commonly used. I'll often see a word, especially a less common words, almost written verbatim in the same exact phrase by multiple authors. This can also help with learning collocations (common pairing of words) too.

Kenkyusha's New Japanese-English Dictionary/新和英大辞典 is also really useful since it has so many example sentences per word, and it's useful for getting a more nuanced feel for how a particular word is used.

For instance, ぼかす. Lots of sentences, but they’re all basically useless. Most seem to be excerpts from technical manuals.)

Funny that you mentioned ぼかす・暈す. While I have seen ぼかすused in fiction, the first few times I saw the word was actually in technical contexts, so that particular does appear in those situations quite often, like when I switched my phone to Japanese and used various camera apps. The very first time I saw the word was when playing a horror visual novel (沙耶の唄 / Saya no Uta). As soon as you start it up, it presents you with settings options screen, asking you the amount of blur you want to apply to the all violent imagery used in the game.

写真のふちをぼかしてみましょう。

クロテスクな画像はそのまま表示
クロテスクな画像のフォーカスを暈かす
クロテスクな画像の明度を落としてフォーカスを暈かす

I've never forgotten the word because of that.

But yeah I agree, when looking for example sentences, in general, I tend to skip over the technical sentences.

I didn't know about the kanji restriction but that makes sense.

As far as readings, yeah, you'll definitely see some wacky ones. There's also a lot names that use ateji and spell things phonetically with kanji like 真理亜 (Maria).

Yeah, it's uncommon, but I've seen that そら reading used in other places too, but it's mainly been in manga, anime, videogames, like in your example.

I wonder if it was a relatively recent fad to read it that way (like in the past 20 years or so?)

In the visual novel 9-episode, there's a kind of a meta comment about that unusual reading.

One of the characters in the visual novel is a teenage girl who was born with the 天 name and そら reading. She intensely hates this name because because everyone usually just calls her てん instead of そら. She describes her name as a 厨二ネーム and she blames her parents for trying to be too edgy, like a 厨二病 teenager edge-lord who has read too much manga and comes up with cringe-worthy names thinking they are cool-sounding but they are not.

Because of that 厨二ネーム, she thinks that's the reason why she's grown up with an eccentric personality, and does strange goofy things, all because of the edge-lord-sounding name. She vows that when she becomes an adult, she'll change her name to something more conventional.

The 9-episode visual novel came out a few years ago, so the girl's parents might have grown up reading manga like Bleach. so I wondered if this was a kind of meta-joke on chuuni-souding names with weird readings.

I laughed when I read the teenage girl's rant about her name because 厨二病 is something you usually attach to teenagers, but here, the teenager is criticizing her adult parents for being like that, so it's a kind of role-reversal.

I read on a Kindle Paperwhite with these custom third-party dictionaries installed.

I also read on my iPhone and iPad using Immersion Reader with Yomichan-formatted dictionaries installed (I use a ton of dictionaries for better word coverage, both J-E and monolingual dictionaries.

The advantage of using Immersion Reader is that you just simply touch the word, instead of selecting and highlighting it, so look-ups are quicker. Also in addition to installing multiple dictionaries, you can install frequency lists, pitch accent info, grammar guides, etc. When you look up a word, it'll search through everything you've install, so if you have pitch accent info and frequency lists installed, you can see the pitch accent of the word you touched, including the frequency number to help determine if it's a word you want to learn.

Then you can save the word to your word list (along with the definition and the sentence containing the word) and later export to Anki.

There also also third party tools that let you export your words from Kindle devices to Anki as well.

Note: some words with furigana may still be difficult to look up correctly on the Kindle and also Immersion Reader, if the formatting of the furigana is weird (a combination of the HTML code and the CSS styling). Sometimes the furigana isn't clearly separated from the word, but instead jumbled together, so the dictionary may not be able to find the word.

To get around this, Immersion Reader has a search function that let you paste in the word (provided you copied it first), so you can fix the spelling (usually it means removing the furigana from the word) and it'll search through all the dictionaries so you can add it to your word list.

Usually kanji-compound words are fine, but some words with kunyomi reading with a single furigana over it may cause some issues, but it depends on the book, and also it may depend on the particular word you are trying to look up.

I read on a Kindle Paperwhite with these custom third-party dictionaries installed.

I also read on my iPhone and iPad using Immersion Reader with Yomichan-formatted dictionaries installed (I use a ton of dictionaries for better word coverage, both J-E and J-J monolingual dictionaries).

The advantage of using Immersion Reader is that you just simply touch the word, instead of selecting and highlighting it, so look-ups are quicker. Also in addition to installing multiple dictionaries, you can install frequency lists, pitch accent info, grammar guides, etc. When you look up a word, it'll search through everything you've installed, so if you have pitch accent info and frequency lists installed, you can see the pitch accent of the word you touched, including the frequency number to help determine if it's a word you want to learn.

Then you can save the word to your word list (it'll also automatically save the definition and the sentence containing the word) so you can later export to Anki.

There also also third party tools that let you export your words from Kindle devices to Anki as well.

Note: some words with furigana may still be difficult to look up correctly on the Kindle and also Immersion Reader, if the formatting of the furigana is weird (a combination of the HTML code and the CSS styling). Sometimes the furigana isn't clearly separated from the word, but instead jumbled together, so the dictionary may not be able to find the word.

Usually kanji-compound words are fine, but some words with kunyomi reading with a single furigana over it may cause some issues, but it depends on the book, and also it may depend on the particular word you are trying to look up.

To get around this, Immersion Reader has a search function that let you paste in the word (provided you copied it first), so you can fix the spelling (usually it means removing the furigana from the word) and it'll search through all the dictionaries so you can add it to your word list. You can also edit each entry in your word list in order to manually add the example sentence, or to remove any unwanted dictionary entries.