Skip to content
EXCAR
Home Product About Roadmap Investors Careers
Join waitlist
Home Product About Roadmap Investors Careers Join waitlist
Build journal

How we built a talking car.

Every session of the EXCAR build, written down: architecture, voice, memory, the app, and the decisions behind them.

00

Overview & Architecture

What EXCAR is, the design philosophy, and the interfaces that make everything swappable.

ArchitecturePhilosophy
01

The Walking Skeleton

The streaming voice loop (Whisper → Claude → Piper), tool use, and wake mode.

VoiceCloud
02

The Offline / Hybrid Brain

Local GGUF + hybrid engine, BT tether, persistent memory, the REST bridge, device boot.

OfflineMemory
03

Mobile App Foundation

The Expo / React Native app: design system, navigation, stores, onboarding, device setup.

Mobile app
04

Dev Tools & Control Surface

An on-device terminal over /exec, the draggable sphere launcher, polling robustness.

Dev tools
05

Fuel, Enrollment & Builds

Fuel & range tracking, voice enrollment (d-vector), EAS build profiles, API contracts.

FuelBuilds
06

Perception & Services

GPS via gpsd, Spotify OAuth, barge-in, audio ducking, diarization, intent gate, Pi config.

PerceptionSpotify
07

Voice Attention & Memory

Echo guard, barge-in hardening, the addressee detector, engine-off memory summarisation.

VoiceMemory
08

Typed Memory Store & App Icon

A provenance + confidence fact store, people/patterns scaffolds, the EX app icon.

Memory
09

Spotify in the App

Structured now-playing + control endpoints, Connect Spotify, the Home now-playing widget.

SpotifyMobile app

No entries match. Try another word.

EXCAR

An AI voice companion for every car on the road.

Instagram info@excar.ai
Site
  • Home
  • Product
  • About
  • Roadmap
  • Investors
  • Careers
Product
  • Hardware & specs
  • Companion app
  • What it does
  • Build journal
Company
  • Our story
  • Team
  • Join our team
  • Investors
  • Contact

© 2026 EXCAR

Privacy · Terms · Contact
PROTOTYPE STATUS: ON THE ROAD