A local-first AI agent with persistent memory, multiple personality modes, and smart backend detection. Works with Groq, Gemini, LM Studio, and Ollama.
Not a wrapper. Not a chatbot. An agent that remembers, adapts, and runs the way you want it to.
Conversations are summarized and retained between sessions. Ellibria remembers context without you repeating yourself.
Switch between personalities in one click — or write your own system prompt from scratch. She adapts completely.
Run locally with LM Studio or Ollama. API keys stay on your machine. No telemetry, no accounts required.
Auto-detects available backends at startup. Local model running? Uses it. Fallback to cloud if not — no config needed.
Natural TTS with female voice selection and language support. Responses read aloud, hands-free.
Follows your Windows system theme automatically. Dark, light, terminal, night — switches in real time.
Actual responses vary based on your system prompt and chosen model.
Auto-detected at startup. No manual configuration required.
Best for privacy. Full local inference, no internet required.
Lightweight alternative. Fast local models, easy setup.
Extremely fast inference. Free tier available at console.groq.com.
Google's fast model. Good fallback when Groq limits hit.
Windows installer is the easiest path. Source install works on Linux too.
Option A — Windows Installer
Download, run, enter your Groq API key. Done.
Option B — From Source
Enter your Groq API key when prompted, or connect a local model.
Get a free Groq API key at console.groq.com/keys