Cody Hurst
Software built with accessibility first.
Software built with accessibility first — using AI to solve problems nothing else on the market is solving.
About
Hi, I’m Cody Hurst.
I’m a technologist based in St. Louis working at the intersection of accessibility, advocacy, and problems nobody else is solving. I’m blind, I use AI to program, and the programming is a means to an end — the end being software that works the way I’d actually want to use it.
My hands-on experience is with HTML, CSS, and Python; the C# behind my flagship applications is written with AI assistance. Where I have real depth is in accessibility itself: WCAG and ARIA review, hands-on work with JAWS, NVDA, VoiceOver, and TalkBack, refreshable braille displays and embossers, and braille translation in Duxbury. I graduated in 2020 with associate degrees in network systems and computer networking, and while the CompTIA and Cisco certifications from that era have since expired, the underlying technical fluency — building my own machines in the mid-2000s, servicing my own Clevo laptops, reasoning through unfamiliar problems — still drives how I work. These days I daily-drive a Framework laptop running a custom Gentoo install, which is an unusual choice among blind users — most land on Fedora, Ubuntu, or Arch. I run Gentoo for the flexibility: I build the system I actually want rather than working around someone else’s defaults.
Accessibility isn’t a skill I picked up — it’s been a necessity my whole life. One moment that stuck with me: in 2017 I helped a blind student model a 3D-printable iPhone stand in OpenSCAD using spatial reasoning and a few lines of syntax. Watching what companies like Orbit Research are doing to bring the cost of refreshable braille down to something ordinary people can afford is a constant reminder that the technology exists — it just has to reach the people who need it.
Outside of that, I write on blog.codyhurst.com and run a community on forum.codyhurst.com. These days I work for myself — independent programming projects, and building a software company around Accessible Trade Terminal. Alongside that I’m open to: accessibility reviews, WCAG audits, AT testing with a user who actually uses the technology, and collaborations where accessibility is a first-class concern rather than an afterthought. If that sounds like a fit, get in touch.
Skills & Credentials
- Using AI to ship real software. I drive Claude Code and similar tools to produce production-grade applications in languages I don’t write by hand — directing architecture, reviewing output, and keeping accessibility as the non-negotiable constraint.
- Screen readers as a daily driver. JAWS, NVDA, VoiceOver (macOS / iOS), TalkBack, and Narrator — used, not just tested.
- WCAG & accessibility review. Evaluating websites and applications for WCAG compliance and real-world usability with assistive tech, and writing up findings that are actionable rather than checkbox-only.
- Assistive technology. Refreshable braille displays, braille embossers, and Duxbury Braille Translator.
- Accessibility-first software design. Building interfaces that reach parity for blind and low-vision users — keyboard-first navigation, ARIA live regions, spatial audio, and semantic narration of otherwise-visual data.
- Web development. Hands-on with HTML, CSS, and Python; comfortable in the MariaDB terminal for everyday tasks; familiar with SQL table structure and relationships.
- Technical support & problem solving. Building PCs, servicing my own hardware, and reasoning through unfamiliar problems using a broad technical background — plus research when research is what’s needed.
- Education. Associate degrees in Network Systems and Computer Networking (2020).
Applications
Desktop applications built with accessibility as a first-class concern — keyboard-first, screen-reader-native, and designed so blind and low-vision users reach parity with sighted users, not a reduced experience.
-
Accessible Trade Terminal
A professional trading and analytics platform built for blind and low-vision traders. It treats market data as a soundscape — spatial panning, frequency-mapped trends, and audio textures let you feel market structure, volatility, and momentum without sight. 26 data providers across crypto, equities, forex, commodities, and macroeconomics.
- Custom 64-voice polyphonic C# DSP engine for real-time sonification — no OS MIDI overhead.
- Native screen-reader speech (NVDA, JAWS, VoiceOver, TalkBack, Narrator) via ARIA live regions.
- Keyboard-first navigation on Windows, macOS, Android, and iOS, with a built-in Help dialog (Alt+H).
- User-buildable Strategy Composer with R:R-gated setups, warmup-aware backtester, and AI Analyst review.
- OS-sandboxed Roslyn custom-indicator runtime (Windows AppContainer / macOS sandbox-exec / Android isolatedProcess) with a security-hardened plugin trust policy.
- 264 tests passing across four target frameworks.
Visit accessibletrader.com (opens in new tab) Try the live demo (opens in new tab)
-
AURA — Accessible User Reading Assistant
A free, open-source screen reader for Windows, written from scratch rather than inherited. NVDA has carried blind computing for twenty years and it is showing its age — it was designed in 2006 and still carries support for technologies that no longer exist. AURA targets the one accessibility API Windows actually uses today, on a modern runtime, with extensibility built in from the start instead of bolted on later.
- Announces focus, state, and position for buttons, checkboxes, lists, menus, and trees — including item counts in older shells that don’t publish them.
- Review cursor for reading by line, word, and character independently of the caret, plus say-all and configurable typing echo.
- Rewritable speech output with no code: how AURA describes any control lives in a plain text file you can edit.
- Versioned plugin contract and project template shipped before 1.0, so add-ons don’t break on every update.
- SAPI 5 and eSpeak NG synthesizers, with live preview of voice, rate, pitch, and volume as you change them.
- Full key remapping, adjustable verbosity, punctuation levels, an installer, and a portable build.
Early development. No browse mode, no braille, no console support, and nobody has used it for real work yet. If you try it, keep NVDA or Narrator running. The repository documents exactly what works and what doesn’t.
-
Accessible Weather Center
A speech- and keyboard-driven weather application with a Weatherscan-style cycling display. Every visual — including radar, normally inaccessible to screen readers — is mirrored by a semantic narration, so blind and low-vision users get life-safety weather information at the same fidelity as sighted viewers, as data and not pixels.
- Ten era-authentic TWC hardware themes (WeatherStar 3000, 4000 v1/v2, Jr, XL, Weatherscan Local/V1/V2, IntelliStar 1/2 HD).
- Accessible map navigation with four lenses over the same spatial data: Places, Alerts, Storms, and Grid Explorer.
- Live NOAA Weather Radio streaming over weatherUSA Icecast feeds, with retry, failover, and its own audio bus.
- Three-bus audio mixer (music / voice / radio) with ducking so narration never fights background music.
- Speaks through NVDA, JAWS, Narrator, Orca or VoiceOver via ARIA live regions. It ships no built-in text-to-speech by design, so your voice, rate and braille display all apply.
- Honors
prefers-reduced-motionand offers a high-contrast overlay on top of any theme. - Runs in the browser with nothing to install, and as a desktop app on Windows, macOS and Linux.
Visit weather.codyhurst.com (opens in new tab) Launch the app (opens in new tab) Repository (opens in new tab)
-
OpenFPS
An audio-first, fully customizable multiplayer game framework for blind and visually impaired players. There are no graphics — gameplay is carried entirely by a binaural soundscape and spatial awareness, so blind players aren’t adapting to a sighted game, they’re playing a game built for them from the ground up.
- HRTF binaural rendering via FMOD Studio with Google Resonance Audio, plus dynamic atmospheric absorption and diffraction so geometry and materials actually sound different.
- Direct NVDA speech bridge through Tolk, with SAPI fallback for users without a screen reader installed.
- Authoritative server with client-side prediction and server reconciliation over reliable UDP — no movement jitter, deterministic physics parity between client and server.
- Zero-allocation hot paths (
ArrayPool<T>,ReadOnlySpan<T>) and a uniform-grid spatial partition for O(1) collision and acoustic queries. - Lock-free, high-priority audio thread for glitch-free binaural rendering independent of game-loop timing.
- Behavior-tree NPC AI and a data-oriented ECS designed so creators can build maps, items, and NPCs end-to-end.
Latest Writing
Community
Discussion, accessibility shop-talk, and project coordination happens on the forum.
Support This Work
If my work is useful to you and you’d like to support it, the simplest way is Cash App:
$churst90 on Cash App (opens in new tab)
Crypto is welcome too, at any of the addresses below. Copy carefully — double-check before sending.
- Ethereum / Arbitrum / Base
0x23b99484A369F718F371c8BFCa5F3FFc53B14F47- BNB Smart Chain
0x23b99484A369F718F371c8BFCa5F3FFc53B14F47- Bitcoin
bc1qp20rjw6vlddmt808v0h8gdrdcmx8n83pmzlaww- Kaspa
kaspa:qypxfclrz0eyw4rn736mwr3ndyfk6c474py9l9mf6n9x9qhpchy6w3cr340mgzy- XRP
rLAaVmjsofxnFJ6o6ZYF3hn6FddKtJnu8q- Solana
HG5BvLsfAgyHGxJe8cfKBzf4qArAmb7LEc5TDe45zwEe- Bittensor (via Tangem)
5GhBmEmMzD92j1rAPMjLbdTmCHQPBZNvyVS2q4RLraEk14jt