DEV Community: Alex Tomita

SayFight - A Real-Time Voice-Controlled Party Game

Alex Tomita — Mon, 28 Jul 2025 01:36:27 +0000

This is a submission for the AssemblyAI Voice Agents Challenge

What I Built

SayFight is a chaotic, real-time, multiplayer party game where players shout funny 3-word phrases to activate their character's actions in fast-paced mini-games. The game uses AssemblyAI's Universal-Streaming API to transcribe speech in real time, allowing players to say phrases like “banana punch blast” or “snake sneak roll” to score points or move their team forward.

This project is built for the Real-Time Performance prompt, showcasing how AssemblyAI enables lightning-fast, voice-driven interactions for games and other real-time apps.

⚠️ Note: This is a proof of concept. Due to time constraints, the project focuses on core functionality:

Real-time voice command recognition

Phrase-to-action triggering

Multiplayer game loop

AssemblyAI integration via WebSocket streaming

The current version includes two simple game modes, each designed to demonstrate the real-time voice interaction loop. While limited in scope, SayFight lays the foundation for a more fully featured voice-controlled party game in future iterations.

Demo

🎮 Live Demo: https://sayfight.com

📸 Screenshots:

GitHub Repository

🔗 https://github.com/rezelco/sayfight.com

Technical Implementation & AssemblyAI Integration

SayFight uses AssemblyAI's Universal Streaming API to transcribe player voice input with low latency. The entire gameplay loop is built around fast transcription, text matching, and synchronized game logic.

🛠 Stack

Frontend: Next.js 14 (App Router) with Tailwind CSS + shadcn/ui, HTML5 Canvas for game visuals
Backend: Node.js + TypeScript with Socket.io for real-time multiplayer
Voice Capture: Web Audio API with MediaStream (not WebRTC – direct audio capture)
Voice Streaming: AssemblyAI Universal Streaming WebSocket API
Audio Processing: 16kHz PCM audio format

🔁 Voice-to-Action Flow

Player joins room and grants microphone permission
- Browser captures audio using MediaStream API
- Audio is processed at 16kHz sample rate with echo cancellation
Audio is streamed to backend via Socket.io
- PCM audio chunks sent as ArrayBuffer
- Not using WebRTC — direct Socket.io binary transport
Backend relays audio to AssemblyAI
- Each player has a dedicated AssemblyAI WebSocket session
- Audio is forwarded in real-time to AssemblyAI’s streaming endpoint
Transcribed text is received with confidence scores
- AssemblyAI returns partial and final transcripts
- Only final transcripts with >0.7 confidence are processed
Command extraction and matching
- Text is matched against the player’s assigned phrase (e.g., “pandas hack wifi”)
- Scoring system:
  - Full phrase match: 2× points
  - Partial phrase: 1.5× points
  - Trigger word only: 1× point
- Total latency target: ~300–400ms
Game state update and broadcast
- Server updates authoritative game state at 60 FPS
- State changes broadcast to all clients via Socket.io
- No client-side prediction needed due to voice command delay envelope

🎯 Key Technical Features

🐼 Phrase-based matching: 100+ unique animal-themed phrases with homophone avoidance
🏆 Multi-tier scoring: Rewards for full phrase vs partial or single-word triggers
🧠 Debug visualization: Host UI shows expected vs received transcription for tuning
🔄 Persistent rooms: Players can reconnect and resume if disconnected
👥 Team-based audio: Tug-of-war mode assigns unique phrases to each team member

🛠️ Lessons Learned

Streaming audio reliably is tricky: WebRTC wasn’t ideal for this use case—direct capture with Socket.io worked better for latency and control.
Phrase design matters: Carefully crafted, phonetically distinct phrases dramatically improved transcription accuracy.
Fast > perfect: A snappy game loop with ~300ms latency was more fun than waiting for ultra-precise recognition.
AssemblyAI impressed: The API was reliable, fast, and handled noisy party environments better than expected.

🚀 Future Plans

If I continue this project, here’s what’s next:

🎮 More game modes like rapid-fire trivia, fighting, or co-op modes
🧠 Flexible NLU with fuzzy matching and intent-based phrase recognition
🌍 Multilingual support via AssemblyAI’s language capabilities
🧩 User-generated content: Let players design their own phrases and game types
📱 Mobile-first UI/UX with animated avatars, emotes, and voice indicators
🧪 Noise resilience and support for simultaneous voice inputs across teams

Thanks to AssemblyAI and DEV for hosting this challenge — building this was a blast!

PurifyPDF – A Privacy-First PDF Sanitizer in Your Inbox

Alex Tomita — Mon, 09 Jun 2025 04:35:14 +0000

This is a submission for the Postmark Challenge: Inbox Innovators.

What I Built

PurifyPDF is a privacy-first document sanitizer that lives in your inbox. Users simply email a PDF to clean@purifypdf.xyz, and moments later receive a cleaned version — stripped of hidden metadata, embedded scripts, and potentially unsafe content.

There’s no app to install, no interface to learn — just send, wait, receive.

PurifyPDF does not retain files longer than necessary. Uploaded documents are stored temporarily for processing and purged after the retention period. User metadata is stored only for usage tracking and verification.

The service prioritizes security by design: PDFs are converted into image-only documents before being rebuilt, ensuring that any embedded scripts, file attachments, hyperlinks, or other potentially harmful elements are completely removed.

PurifyPDF is ideal for:

Privacy-conscious professionals
Journalists, lawyers, or auditors
Anyone sharing or receiving documents from unknown sources

Demo

To test the service:

Send an email to clean@purifypdf.xyz with the subject: register
You’ll receive a one-time verification link. Click it to verify your email
Send another email to clean@purifypdf.xyz with a PDF attached
Wait about a minute and receive a sanitized version in your inbox

PDF size limit: 5 MB

Recommended page limit: 25 pages

Usage limit: Each verified user can sanitize up to 3 PDFs

Judges: If you need your usage limit increased for evaluation, please email support@rezelco.com.

Code Repository

GitHub: https://github.com/rezelco/purifypdf-n8n

rezelco / purifypdf-n8n

Email-based PDF sanitizer using Postmark, PDF.co, Airtable, and n8n

PurifyPDF – Email-Based PDF Sanitizer

PurifyPDF is a privacy-first PDF sanitization workflow built using n8n, Postmark, PDF.co, and Airtable.

Users send a PDF as an email attachment to clean@yourdomain.com, and the system returns a sanitized version, free of metadata, embedded scripts, and potentially unsafe content.

🚀 Demo

Send an email to clean@yourdomain.com with the subject: register
You’ll receive a verification link. Click to activate your email
Send another email with a PDF attached
Receive a sanitized version in your inbox

Limits

Max PDF size: 5 MB
Recommended page count: 25
Each user may sanitize up to 3 files

🧰 Tech Stack

n8n – Orchestrates workflow logic
Postmark – Handles incoming and outgoing emails
PDF.co – Converts and rebuilds PDFs for sanitization
Airtable – Tracks users, limits, and verification

🔒 Privacy

PurifyPDF never stores files permanently. All PDFs are purged automatically after one hour. Only metadata…

View on GitHub

This repository includes:

The full sanitized n8n workflow (purifypdf.json)
MIT License and setup documentation

See the README for installation steps, required credentials, and detailed architecture.

How I Built It

PurifyPDF is built entirely in n8n, using no-code and low-code logic to orchestrate document processing, verification, and email flows.

The goal was to create something that feels invisible to users — no apps, no dashboards — just smart automation behind a simple email address.

Tech Stack

n8n – Manages the full automation pipeline, including webhook ingestion, token generation, user tracking, and email sending.
Postmark – Powers both inbound and outbound email. Incoming attachments are parsed through webhooks, and outgoing messages are branded, responsive, and fast.
PDF.co – Sanitizes documents by converting each page to an image and rebuilding a new, clean PDF — free of metadata, scripts, and embedded content.
Airtable – Stores user records, verification tokens, timestamps, and usage limits.
JavaScript (Function nodes) – Used for logic branching, token generation, rate limiting, and retry logic inside the n8n flow.

Design Features

Email-first UX – Every interaction happens through email. No UI required.
One-click user registration – Email-based verification with expiring tokens
Retry handling – Long-running PDF conversions use polling + wait loops
Usage limits – Users can sanitize up to 3 PDFs unless upgraded
Security & privacy – Files are stored temporarily and purged after processing

n8n Flow

Why I Chose Postmark

Postmark made it easy to:

Receive emails and attachments via webhooks
Monitor all incoming and outgoing email activity — incredibly helpful for debugging and delivery tracking
Seamlessly integrate with n8n, making it easy to trigger flows, parse messages, and send responses without glue code

Challenge Reflection

This project started with a real, everyday problem at home. I found out my wife keeps a separate laptop that she uses only to open vendor invoices. After opening them, she prints the files, walks them to a second computer, and scans them back in — all to avoid the risk of opening untrusted PDFs on her main machine.

It was a clever workaround, but it got me thinking: in a world full of automation tools, why are we still physically printing documents just to feel safe?

That idea became the foundation for PurifyPDF. I wanted to build something simple, trustworthy, and accessible — a tool that requires no apps, no learning curve, and no new habits. If someone like my wife could use it without friction, that meant it could help a lot of other people too.

The Postmark Challenge was the perfect setting to bring it to life. Postmark made handling inbound and outbound email effortless, and pairing it with n8n and PDF.co gave me the flexibility to build a secure, responsive service without writing a backend from scratch.

This was one of the most rewarding projects I've worked on. It solves a real problem, feels invisible to the user, and showcases just how powerful email-driven automation can be.

License

MIT License

The name PurifyPDF and associated branding (e.g., logo, domain name) are not covered by the MIT license and may not be used without permission.