Ember AI (formerly Inferra AI)

Ember AI — Bonfire Base Studio

A multi-modal AI system spanning desktop, Discord, and the web. Unified memory, real-time inference, one continuous thread.

Three platforms

01 — Desktop overlay Always-on-top floating HUD with local voice recognition and zero background throttling. Single-click to arm, hands-free from there.

02 — Community bot AI-powered Discord assistant handling moderation, GitHub webhook relay, new-member onboarding, and 24/7 uptime via cloud watchdogs.

03 — Web console React + TypeScript dashboard with streaming chat, Cloudflare Turnstile bot protection, and full cloud access from any browser.

Capabilities

• Persistent memory — vector embeddings retain context and preferences across sessions
• Live data retrieval — real-time Web search and PAGASA weather and hazard alerts
• Multi-modal input — switch between voice and text across desktop and web
• Task scheduling — background jobs, reminders, and timed dispatches via cron jobs or lifecycle.
• GitHub integration — commit logs, issues, and PRs surfaced directly into Discord
• Security layer — rate limiting, auth middleware, and Turnstile bot protection, firewall tempering detection.

Stack

Frontend — React, TypeScript, Vite, Cloudflare Turnstile Desktop — Electron, Node.js IPC, Python STT Backend — Node.js, Discord.js, Express, QStash AI — Groq LLaMA, Mixtral, Vector search Data — Firestore, MongoDB, Redis APIs — GitHub, Tavily, 

Ember replaces fragmented single-purpose tools with one continuous intelligent system — context and operational readiness across all the Bonfire Base Studio Ecosystem.