A guided tour of every part of the app

Crosstown is a complete, daily-life learning and safety companion built to help you build and sustain an independent life in New York City, and it was designed from the ground up for people with ADHD, visuospatial learning disabilities, blindness, dyspraxia, cerebral palsy, autism, Down syndrome, and vestibular conditions, though absolutely anyone can use it and benefit from it. The app is organized around two guiding characters who stay with you throughout everything you do here — Dr. Mira, your psychiatrist, who runs the emotional regulation and self-defense courses, talks with you in a full AI-powered chat, and reads every single lesson out loud in a voice that you personally pick and calibrate so that it is comfortable for you to listen to for long stretches, celebrating your progress with real dramatic gasps and genuine praise that makes you feel seen; and Dr. Chen, your testing doctor, who administers the IQ test modeled on the WAIS, runs the Watch & Guess video quizzes, grades your answers, explains each one in plain language, says congratulations after every lesson you pass, and serves as the bridge between you and Dr. Mira by carrying messages back and forth, telling you what she thinks of your progress, and even walking into her office with your printed progress report to set on her desk. The app has six main areas that you can explore in any order, and all of your progress is saved automatically on whatever device you are using, with a six-character sync code that lets you copy your completed lessons between your phone and your computer so that they always match and nothing is ever erased — the sync runs automatically on load, whenever the tab comes to the front, and every sixty seconds in the background. The first main area is the Street Safety Courses, a year-long daily curriculum of three hundred and eighty-eight lessons, each one a short three-to-five-minute read covering a real NYC topic like reading the WALK signal, making eye contact with turning drivers, standing back from subway platform edges, understanding bike lane flow, tapping in with OMNY, choosing the right late-night subway car, and dozens more, with a quiz after each lesson that marks it complete after a single correct attempt and a green check that never disappears, and when you finish all three hundred and eighty-eight the app celebrates automatically with Dr. Chen shouting WOW and Dr. Mira sitting down in shock. The second area is Watch & Guess, where each day has a short video clip of a real NYC street scene that you watch and then answer thirty questions about which behavior is safe, which is risky, and why it matters, with clips that preload and chain together so the video never freezes, and Dr. Mira reads each lesson out loud in her calibrated voice with an animated mouth that moves as she speaks. The third area is the Self-Defense Course, fifty videos across five modules — Awareness & Positioning, where you learn to scan the whole scene before helping anyone because a second person hurt is a worse problem not a better one; Voice & De-escalation, where your voice is your first tool; Hands-On Basics, simple physical skills that work under stress; Subway & Night Travel, covering platform safety and car choice; and Emotional Regulation Under Pressure, covering breathing, freezing, and recovery — each lesson with its own video, summary, tips from Dr. Mira, and an audio read-aloud. The fourth area is the Emotional Regulation Course, sixty videos across sixty modules — one focused module for every lesson, organized around Foundations, Body Awareness, Mind & Thoughts, Relationships, and Daily Practice — where every lesson shows Dr. Mira in her office in her black blazer with her hair down and glasses on, typing up your notes, opening a drawer, and flipping through your pages, and you can pick between two background songs both pitch-matched to the same musical key so they sound harmonious together, and she reads each lesson aloud in the voice you calibrated. The fifth area is the Transit Tools & Quizzes, with live subway and bus arrivals shown on the home page, a Transit Navigator quiz that tests your knowledge in eight questions, a How to Use Public Transportation hub with twelve step-by-step guides from tapping in with OMNY to taking the Staten Island Ferry, and an Answer Key page listing every lesson's correct answers and explanations. The sixth area is the Accessibility Toolkits, six specialized tools each designed for a specific condition — the Dyspraxia toolkit for straight-path and curb-step practice, the Cerebral Palsy crossing aid with extended timing and paced steps, the Autism traffic sense module for predicting speed and pausing before crossing, the Vestibular crossing aid with gaze anchors and dizziness-safe steps, the Down Syndrome crossing aid with picture cues and step checklists, and the Blind traveler support with audio narration and audible signal guidance. Beyond those six areas, the app also gives you a full chat with Dr. Mira powered by AI for help with ADHD, anxiety, overwhelm, and emotional regulation; a complete thirty-question IQ test that Dr. Chen administers and that he reads the reading-comprehension questions out loud for because Dr. Mira said he could, with results forwarded to her for scoring; a "make writing stronger" tool that polishes your written explanations with an Undo button; an animated office scene where you can watch Dr. Mira walk through her fifth-floor corner office with her legs stepping in alternating strides, then sit at her desk and type, open a drawer, and flip through your pages; voice calibration and gasp settings where you pick Dr. Mira's tone, pitch, speed, and how dramatic her gasps are, and once saved the same voice plays everywhere in the app automatically; background music with several original Suno tracks all pitch-matched to the same key, with volume, bass boost, and voice profile controls that save automatically and work the same on iPhone and computer; a progress dashboard that shows your total progress across every course in one place with no syncing needed; Dr. Chen's story cards scattered across the home page with short voice notes about what Dr. Mira thinks of your progress, including a perfume quiz where Dr. Mira reacts to your favorite scent in her own voice; and milestone celebrations that replay automatically when you complete all three hundred and eighty-eight lessons. Everything here was built so that you can start anywhere that interests you, go at your own pace, and know that your two guides are always paying attention to what you accomplish and ready to celebrate it with you.

Under the surface, Crosstown is a single-page application built on TanStack Start with React and Vite, which means the entire interface runs in your browser and updates instantly without reloading the page, while the heavier work — generating Dr. Mira's spoken voice with real dramatic gasps, polishing your writing, scoring the IQ test, and relaying Dr. Chen's messages — happens on the server through typed server functions that call the Lovable AI Gateway so no API keys are ever exposed to your device. Your completed lessons, voice settings, music preferences, perfume answers, and quiz picks are stored locally in your browser's localStorage, so the app keeps working even with no connection, and they are also written to a cloud database behind Supabase row-level security, which means only your own account can read or change your data — the six-character sync code is simply a share key that tells the server to merge your phone's progress and your computer's progress into one record without ever deleting a completed lesson. Dr. Mira's voice is produced in two complementary ways: for her course read-alouds, office narration, and milestone reactions, the app uses your device's built-in speech synthesis engine with the exact voice, pitch, and speed you calibrated, chunking long scripts on sentence boundaries and holding each utterance in memory so iOS does not garbage-collect it mid-sentence; for her story-card reactions and Dr. Chen's messages, audio is streamed from the server-side TTS endpoint, which inserts gasp cues — sharp inhales, shaky breaths, and full gasps at varying frequency and intensity depending on the profile you selected — into the spoken text before synthesis so the emotional coloring sounds natural rather than robotic. The background music runs through a Web Audio graph that applies bass boost and a five-band voice profile filter in real time, with a native-audio fallback on iPhone Safari that avoids the audio-context autoplay restrictions by unlocking on your first tap; all tracks are pitch-matched to the same key at build time so switching between them never sounds jarring. The Watch & Guess videos and the emotional-regulation lesson clips are preloaded and chained with a watchdog timer that detects any frame stall and immediately advances to the next buffered clip, which is why playback never freezes and your completed exercises never disappear even if the network drops mid-lesson.