Show HN: ReadKinetic – a free, local-first speed reader for your own books

Thanks. To be frank, the images and graphs are the weak spot at the moment. The parsers extract just the text layer and that’s the extent of it, so images are just discarded and you don’t even get told there was one.

Tables are worse than dropped - PDF extraction de-flattens them into a stream of values in reading order, so you just get a torrent of disconnected numbers flitting past which don’t mean a great deal.

I would read figures in the normal way for this type of book and then use the tool for the prose. It’s not really fixable at the format level too, there is no good way of showing a person a graph word by word. I suppose what I really should add is some kind of flag to say there was a figure here instead of just letting the process go ahead and drop everything. Otherwise, you can’t really know that you’re missing things.

Thanks. The parsing cost is the entire one-time cost, the rest of the operation involves simply walking an array of words, a 900 page novel and a short story will perform the exact same. There is essentially zero cost in rendering a word at a time no matter the size of the book.

PDF is significantly the worst. We use pdf.js, and it has to page by page to retrieve the text layer, so a large book takes a little while and the tab feels 'busy' while doing so. EPUBs are considerably faster, it's mostly a case of unzipping and parsing xhtml.

The main thing that was the difference maker was separating storage into two IndexedDB stores, one for metadata, and one for the full text. When we open the library list we only read from the metadata store, not pulling the entire content of all books you own into memory. We only pull in the one you actually want to read.

The biggest edge case by far are scanned PDFs that have no text layer (there is no OCR, so nothing is returned) that are currently not failing audibly enough to make it clear.

Cool idea!

Maybe you could somehow integrate it with Calibre [0].

https://calibre-ebook.com/download

Thanks. Calibre’s crossed my mind a few times, the obvious hook is the content server, but the browser integration makes it clunky. Calibre-server doesn’t spit out CORS headers that you can use from another origin, so a web app can’t just read your library over HTTP unless you do a bit of reconfiguring it, and making people set CORS flags just for my little tool seems like a bad bargain.

The avenue that actually looks viable is directory import. Calibre keeps everything as normal files in Author/Title folders with a metadata.opf next to each book, so if I point a directory picker at your library directory, I can traverse it, grab the opf file to get the actual title/author rather than having to guess based on the filename, and then suck the epub files directly in. Would be a properly slick bulk import.

Catch is the File System Access API only works in Chromium at present, so anyone using Firefox/Safari would still be forced to pick their files one by one. Would be a worthwhile trade-off though as it gracefully degrades.

Yeah the toggle is there-double press and it locks it into playing continually. But if you didn't find it, that’s on me because it’s not discoverable. There is that line of text at the top which says hold to read, left/right to scrub, and obviously if that isn’t making clear there is another mode of play lurking just behind a gesture no one ever brings up.

I did default to hold because the read stops the instant you are not reading – if your attention is drifting, words aren't racing on ahead.

Which felt right for an application whose whole purpose is concentration. But you are almost certainly going to react the way most people will, and there should be an easy toggle instead of one you’re only likely to find when you stumble over it by accident. Scrub is already mapped to left and right-up and down would be an excellent match to speed, never thought of it that way until you said it. Let me know how you get on with the longer pieces, that’s where I’d expect either that it just clicks (or fails miserably).