Last month a policy researcher I know counted the PDFs sitting in her “To Read” folder: 63. At an honest 250 words per minute, that’s roughly 19 hours of straight reading, before she opens a single email. She didn’t delete the folder. She started listening to it.
Speechify is the app most people try for exactly this problem, and most people abandon it inside two weeks. Usually it’s not the technology that fails. It’s the setup. They play the sample paragraph, nod approvingly, and never get their actual backlog inside the app. If you want the wider verdict on the tool before you commit an evening to it, there’s a solid Speechify review of what the reader does well and where it stumbles. What follows is the part that review won’t give you: the order to set things up in, and the numbers that make it stick.
Step 1: Put your real reading pile into the app
The sample text is a trap. Nobody has a habit built on sample text. Spend fifteen minutes importing the things you’re genuinely behind on, and the app stops being a demo.
Speechify accepts more formats than most people realise:
- PDFs and EPUBs — upload a file directly, or pull one in from Google Drive, Dropbox or your phone’s storage.
- Web articles — the Chrome extension reads whatever page you’re on and strips out the menus, pop-ups and adverts first.
- Google Docs — handy if you write long-form and want to hear your own drafts read back to you.
- Camera scans — photograph a printed page and the built-in OCR converts it into readable text.
- Shared links — on mobile, the share sheet pushes a newsletter or article straight into your library in two taps.
Then file them. I use three folders: Client, Learning, and Later. The naming matters less than the separation, because the three categories want different playback speeds.
Step 2: Pick one voice and live with it for a week
The voice library is where people burn the most time for the least benefit. Speechify ships hundreds of voices, including celebrity options like Snoop Dogg and Gwyneth Paltrow. Delightful for a fifteen-second clip you send to a friend. Unbearable by chapter nine of a compliance manual.
Choose a neutral, mid-paced English voice and don’t touch it for seven days. Switching voices resets your ear, and a reset ear can’t climb into higher speeds without noticing. Consistency is what eventually makes 2x feel like normal conversation.
If you’re tempted to build something more personal, custom narration is a bigger project than a dropdown menu implies — the process of capturing reference audio and training a model is walkable end to end in this hands-on guide to cloning a voice and generating speech. Worth knowing that the same underlying techniques are also what security teams use to detect cloned voices and deepfake audio. Useful context before you upload your own recordings anywhere.
Step 3: Raise the speed in small increments, not leaps
Two failure modes here. People stay at 1x, get bored, and quit. Or they jump straight to 4x, understand nothing, and quit. The useful path sits in between and takes about ten days.
A progression that works:
- Days 1–2 at 1.25x. It will feel sluggish. That’s fine.
- Days 3–7 at 1.5x.
- Week two: 1.75x, then 2x.
- Most people plateau somewhere between 2.2x and 2.8x on non-fiction and stop there permanently.
The 10 percent rule
Nudge the slider by roughly a tenth at a time, never more. If you catch yourself rewinding to catch a sentence you missed, you’ve overshot. Drop back one notch and stay there for a few days. Chasing a number is pointless; the speed that suits a legal brief is not the speed that suits a memoir. Dense, statistics-heavy material usually reads best at 1.5x to 1.8x. Narrative non-fiction and email hold together above 2x without much effort.
Step 4: Turn your inbox into a listening queue
Newsletters are the highest-return thing to convert, because they’re already written in short, spoken-friendly paragraphs. A 2,000-word newsletter takes about seven minutes at 2x. Three of them fit inside a coffee break that you’d otherwise spend scrolling.
The mechanic is simple. Instead of tapping a link and reading on your phone, share the article to Speechify. It lands in your library, queued, ready for the commute. Six months of doing this and the twenty minutes between the train doors and the office becomes the most productive reading slot in your week.
Step 5: Scan the paper you can’t avoid
Plenty of material still arrives printed or as a scanned image, and a scanned PDF that refuses to highlight is a special kind of frustrating. Use the camera scanner and the text becomes searchable and speakable.
Realistic expectations: a clean single-column page takes two or three seconds each. A 38-page vendor contract is a fifteen-minute job including a quick proofread. Two-column academic layouts and handwriting will produce errors, so budget a cleanup pass for those. The upside is highlight sync — when the narrator reaches something important, you can tap the sentence in the text view and jump straight to that spot for citation.
Step 6: Dictate instead of typing when the words are already in your head
Most people type around 40 words a minute and speak at closer to 130. For first drafts, meeting notes and replies you already know the shape of, dictation closes that gap fast. It is not a substitute for editing. It’s a way of getting a rough version onto the page before the thought evaporates.
One thing worth checking before you rely on it: where the audio goes. The Windows version of Speechify processes transcription and dictation with local models, which means the recording doesn’t have to leave your machine and a flaky connection won’t kill the session. If you dictate client calls or medical notes, that distinction matters more than any voice quality setting.
Step 7: Give it a fixed slot in the week
Tools that live inside a routine survive. Tools that wait for a free moment don’t. This is the pattern I’d copy:
- Monday morning: dump the week’s PDFs and reports into a folder and upload them in one batch. Five minutes.
- Commute: newsletters and anything under 3,000 words.
- Lunch walk: twenty minutes of whatever is longest and least urgent.
- Friday afternoon: dictate the week’s notes and draft replies while tidying your desk.
That schedule adds up to roughly four and a half hours of listening a week without touching your working hours. At 2x, that’s close to nine hours of reading you’d otherwise never have done.
The three habits that quietly kill a listening routine
Hoarding instead of finishing. A library of 200 unstarted articles feels like progress and isn’t. Cap it at ten items and delete anything that’s been sitting for a month. You will not miss it.
Listening to everything. Dense technical material with tables, equations or code needs your eyes, and no playback speed fixes that. Send it to the “Later” folder and read it properly on a screen. Speechify is at its best on prose: articles, reports, contracts, email, book chapters.
Refusing to rewind. Speed only helps if you actually absorb the material, so treat a rewind as information rather than failure. If you’ve rewound the same paragraph three times, that’s your cue to drop the speed by a quarter and slow down for a page or two. The habit compounds when you let it flex.