I want to use as little Google software as I can get away with. This goes as far as things like the Android GBoard on my phone. I want an opensource keyboard, and there are plenty. None of the ones I tried seem to have an emoji search. This sounds really minor, but given how many emojis there are to scroll through, the only way to find them is by search.

FlorisBoard came the closest. It can suggest an emoji when you type its name, for example, typing "confetti" would suggest 🎊. This was OK, but it wasn't great, where it fell down for me was that it only matched on exact strings, typing "smile" only suggested 😼 because ☺️ is "smiling face".

Personal Emoji Dictionary

The solution was to use a "personal dictionary". I could add 😁 (Beaming Face with Smiling Eyes) and set it as a synonym for "grin" and "smile". This dictionary is agnostic of the keyboard app, so I was free to pick based on other criteria. I picked HeliBoard.

The problem now is that there are about 2,000 common emojis... I only use a fairly small number of them, but the idea of manually entering even most of them through the Android UI didn't sound like fun. Time for some automation 🤖

I needed to figure out where the dictionary was stored on the phone. I started with the file explorer app, don't know why I bothered, it's useless. Next I connected via ADB and was able to find it. It's an SQLite DB:

/data/data/com.android.providers.userdictionary/databases/user_dict.db

SELECT * FROM words

1|🎉|250|en_GB|0|tada
2|😁|250|en_GB|0|grin
3|😁|250|en_GB|0|smile

To generate the entries, threw the problem at Claude and got a decent CSV, which I could then import into SQLite. It was a long way from perfect, but it was good enough after some fine-tuning for to my personal taste.

For reference, the locally created DB was copied to the phone using ADB:

adb push emoji_en.dict /data/data/com.android.providers.userdictionary/databases/emoji_en.dict