A little under a week ago I came across a post about RSVP reading — a technique where text flashes one word at a time at a fixed point on the screen, eliminating the eye movement that slows most people down. I forwarded it to some friends on Instagram. Didn't think too much of it at the time.

A few days later I checked in on the group chat. Asked if anyone had actually tried the RSVP demo in the post. We all gave it a shot. After we tried it, I said something like — we should build something like this if it doesn't already exist. Tools like this should be everywhere.

That conversation is what became FlowRead. By the end of the night I had a working v1.2 built from scratch using AI tools. My daughter tested it at 250 words per minute on her first try. That was the moment I knew this was something real.

WHAT IS RSVP READING?

RSVP stands for Rapid Serial Visual Presentation. Instead of reading text the traditional way — scanning left to right across a line, jumping to the next line, losing your place, re-reading — your eyes lock on a single fixed point on the screen. Words appear one at a time, right at that point, at a speed you control.

The result is that all the energy your brain normally spends moving your eyes gets redirected into actually processing what you're reading. No saccades — the technical term for the rapid eye jumps that happen during normal reading. No skipping back. No losing your place. Just pure word-by-word comprehension.

Why Does It Work?

Research into RSVP reading goes back decades. The core finding is simple: eye movement is one of the biggest bottlenecks in reading speed. When you eliminate it, most readers can hit two to three times their normal reading speed without losing comprehension — and in some cases, comprehension actually improves because the fixed pace forces full attention on each word.

FlowRead adds one more layer to this: BPM-synced generative music. The tempo of the background music is mathematically locked to your reading speed. At 250 WPM, the music plays at approximately 95 BPM. Slow down, the music slows. Speed up, it follows. The rhythm anchors your attention and helps your brain settle into a reading flow state — similar to how a metronome helps musicians maintain tempo without thinking about it.

The Science Behind the Sync BPM = WPM × 0.38. This ratio keeps the music tempo in a range that complements cognitive processing without overwhelming it. At typical reading speeds (200–300 WPM), this produces music between 76–114 BPM — the same range as most focus and study playlists.

WHO IT HELPS

RSVP reading is particularly valuable for people who struggle with traditional reading formats. The fixed focal point reduces the cognitive load of eye tracking, which is especially helpful for readers with dyslexia, ADHD, or attention difficulties. The word-by-word pacing also prevents skimming — a habit that feels efficient but often kills comprehension.

But it's not just for people who find reading difficult. Any reader who wants to get through more content — articles, research, long-form writing — in less time will benefit from training with RSVP. The speed is adjustable from 60 WPM all the way to 800 WPM. Most people start around 200–250, find their ceiling through testing, and build from there.

My daughter hit 250 WPM on her first try. For reference, the average adult reads at about 200–250 WPM silently. She was reading at average adult pace without any training. That's what the fixed focal point does — it removes the inefficiency immediately.

200
Avg adult WPM
250
Daughter's first test
800
FlowRead max WPM

HOW WE GOT FROM V1.0 TO V1.2 IN ONE NIGHT

The build happened fast. Here's the honest timeline:

V1.0
The original FlowRead Basic RSVP reader with BPM-synced generative music. Five genres: Jazz, Orchestral, Lo-Fi, Percussion, Silent. ORP (Optimal Recognition Point) highlighting. Punctuation-aware pacing. Single HTML file, zero dependencies. You paste text, it reads it. Clean, functional, but standalone — no connection to the site itself.
V1.1
Site integration attempt Added site content detection, theme mimicry, and font matching. The idea was right — FlowRead should read the site automatically without copy-paste. But the implementation tried to use cross-origin DOM access to read the main site's CSS variables and content from a separate page. That doesn't work. The sections showed "No sections detected." The theme didn't apply. Back to the drawing board — the concept was right, the execution wasn't.
V1.2
The fix — hardcoded and working Scrapped the cross-origin approach. Fetched the live site content directly and hardcoded all seven sections (Intro, Craft, Music, Projects, Services, Articles, Contact) into the tool itself. Hardcoded the Spark1Early black/red/white color system and Bebas Neue font directly.

v1.3 — New genres Expanded from 4 genres to 8. Added Theta Ambient (slow drone pads named after the 4-8Hz focus brainwave), Solo Piano (melodic right hand over walking bass, new arrangement every session), Big Band Swing (punchy brass stabs with triplet swing hihat), and a dedicated Percussion mode with a full layered kit. This is the version live now.

The whole arc — group chat post to live v1.2 — happened in a single night. That's what building with AI actually looks like. You don't get it right on the first try. You iterate fast, test it, find out what's broken, and fix it. Version numbers aren't milestones, they're just timestamps on the work.

THE ORP — WHAT THAT HIGHLIGHTED LETTER DOES

Every word displayed in FlowRead has one letter highlighted in red — the ORP, or Optimal Recognition Point. This is the specific letter your eye naturally anchors to when reading a word. Research shows that the human visual system doesn't read words character by character — it recognizes word shapes, and the ORP is where that recognition happens fastest.

FlowRead calculates the ORP based on word length: one character in for short words, two or three in for longer ones. Text before it right-aligns to center. Text after it extends left. Your eye locks on the same spot for every single word. This is the detail that separates RSVP reading from just flashing words on a screen.

PUNCTUATION PACING — WHY IT BREATHES

A common criticism of RSVP reading is that it feels robotic — words flashing at a fixed rate with no natural rhythm. FlowRead addresses this with a punctuation-aware delay system. Every word gets a custom delay based on what punctuation follows it. A period gets nearly three times the normal delay. An ellipsis gets even more. A comma gets a short pause. A hyphenated word moves slightly faster.

The result is reading that breathes the way spoken language does. Your brain gets the same micro-pauses at sentence boundaries that it's used to from listening and speaking. The words flash faster when momentum builds, slower when a thought needs to land. Combined with the music sync, it stops feeling like a machine and starts feeling like a rhythm.

WHAT'S NEXT — FREEMIUM

FlowRead is not going open source. It's going freemium.

Looking at the competition — Spreeder charges around 7, SwiftRead has a Pro tier, AccelaReader is free with no monetization. None of them have BPM-synced generative music. None of them have 8 procedurally composed genres. That's the differentiator, and open sourcing it gives that away. The plan is to protect it and build it into a sustainable product instead.

The free tier stays genuinely useful. The Pro tier unlocks depth — the premium genres, URL import so you can paste any article link and it reads automatically, PDF support, reading history, and site embed so any publisher or blogger can add FlowRead to their own site with an API key.

Early Access All 8 genres are currently available with no gate while FlowRead is in early access. The freemium model activates at official launch. Try everything now.

Seven sections of this site, ready to read. Pick a speed. Pick a genre. Hit play.

Try FlowRead v1.7 →
A note on the build FlowRead was built entirely using AI tools — Claude for architecture and code generation, iterating through three versions in one session. No frameworks. No build tools. One HTML file. This is what AI consulting looks like from the inside: not using AI to replace thinking, but using it to collapse the time between an idea and a working product.

VERSION HISTORY

FlowRead moved fast. Here's every version and what changed:

V1.0
The original build Basic RSVP reader. BPM-synced generative music via Web Audio API oscillators. 4 genres: Jazz, Orchestral, Lo-Fi, Percussion. ORP highlighting. Punctuation-aware pacing. Single HTML file, zero dependencies. Paste text, hit play.
V1.1
Site integration attempt Added site content detection and theme mimicry. Cross-origin DOM approach failed on standalone pages — sections showed "not detected," theme didn't apply. Right concept, wrong implementation.
V1.2
Hardcoded and working Scrapped cross-origin approach. Hardcoded all 7 site sections and Spark1Early theme directly. Section nav works instantly. Deployed live at spark1early.github.io/tools/flowread.html.
V1.3
8 genres Expanded from 4 to 8 genres. Added Theta Ambient (named after the 4–8Hz focus brainwave), Solo Piano, Big Band Swing, and a dedicated Percussion mode. Freemium model confirmed — not open source.
V1.4
Better synthesis New instrument-specific synth functions: piano(), bass(), brass(), pad(), drone(). Swing jazz hihat fix with proper triplet t-tt-t-tt grid. Article page integration via ?article= URL parameter.
V1.5
Tone.js upgrade (rolled back) Attempted upgrade to Tone.js for realistic instrument synthesis — PolySynth piano, MembraneSynth drums, MetalSynth cymbals, Reverb/Chorus effects chains. Mobile AudioContext suspension issues caused music to stop on genre switch and WPM change. Rolled back.
V1.6
Back to Web Audio — melodic sequences Returned to stateless Web Audio API. Replaced random notes with fixed melodic sequences per genre. Added MELODIES, BASSLINES, and CHORDS tables. Theta Ambient fixed — faster attack so it's actually audible. Genre switching instant with no transport restart.
V1.7
Real recorded audio loops — current version Replaced all synthesized instruments with real recorded loops from freesound.org (CC0 licensed). Jazz sounds like jazz. Lo-fi sounds like lo-fi. BPM sync via Web Audio API playbackRate — tempo adjusts live as you move the WPM slider. Pre-loads all 7 genres in background after first play for instant switching. Works offline after first load. 7 genres: Swing Jazz, Lo-Fi, Orchestral, Theta Ambient, Solo Piano, Big Band Swing, Percussion.

— Spark1Early · Boston, MA · 2026