AI summary by TwinMind

My new app is here. It's really fast.

Theo - t3․gg

11m 8s

App Launch Announcement

  • Launched T3 Chat - fastest AI chat app built
  • Free tier available, Pro tier $8/month for higher message limits
  • Over 100,000 messages sent in first 12 hours before video release
  • Built in 5 days with unique architecture

Speed & Performance Focus

  • Emphasizes speed over AI features - "AI is not what makes this interesting"
  • Real-time streaming without editing for demo
  • Local-first architecture - works offline for navigation
  • Can search across message history locally
  • No loading states, only stale states for better UX

Technical Architecture Deep Dive

  • Furthest departure from T3 stack despite name
  • Built static app shell using Next.js with custom routing
  • React Router app forced into Next.js framework
  • SSR disabled (SSR: false) for entirely client-side operation
  • Custom rewrite rules in next.config redirecting to static app shell

Database & Storage Implementation

  • Uses Dexie library as wrapper for IndexDB browser storage
  • Built custom sync layer instead of using Dexie Cloud
  • Combines local storage, data fetching, and caching
  • Auth state stored locally as backup, not single source of truth
  • Hacks Dexie extensively for sync layer experience

Performance Optimizations

  • Custom link component using onMouseDown for navigation
  • Prefetches everything possible
  • Abuses React compiler and memoization
  • Collaborated with Aiden (Million.js creator) for React performance
  • Used React Scan for performance analysis

Model Selection Update

  • Added model selection feature after user feedback
  • No additional cost - still $8/month for 500 messages per 7 days
  • Pricing model may change with email notifications
  • Willing to absorb costs while figuring out optimal pricing

Competitive Landscape Discussion

  • Acknowledges product could be obsoleted if competitors improve
  • Compares to previous products (Ping, Upload Thing, etc.)
  • Notes ChatGPT team busy rewriting Next app in Remix
  • Built for better experience, not to monopolize market
  • Expects copying attempts but confident in execution difficulty

Planned Features

  • Dedicated desktop app (high priority)
  • File uploads
  • Projects with context
  • Local search functionality
  • Branching chat conversations
  • Hot keys and desktop behaviors

Community Engagement

  • T3 Chat Discord channel created for feedback
  • Community-driven development approach
  • Will prioritize user-requested features over planned roadmap
  • Wants to be receptive to user needs and problems

TwinMind

Get summaries like this for your own meetings

TwinMind takes notes in the background during meetings, lectures, and calls, then writes the summary for you.

Free plan to get started

No bot joins your calls

Phone, desktop & Chrome apps

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome

Your next meeting, summarized like this

TwinMind listens in the background and turns meetings, lectures, and calls into searchable notes you can chat with. Start free.

Available on phone, desktop & Chrome