Early access · 2026

AI companions that talk back, live.

Soomiles builds Animalive: characters that read the room. They follow live chat, pick what to answer, and reply out loud in under three seconds — with a voice, a face, and a personality that stays consistent.

Get in touch

viewer_42: Haru, it's raining here. What are you drinking?

Haru: Hot cocoa, obviously! Rainy days are the best for it. Stay dry out there, okay?

What we're building

Real-time conversation

An LLM picks and answers chat messages as they arrive, prioritizing direct questions and supporters.

Voice and expression

Streaming text-to-speech with lip sync, and Live2D expressions that follow the mood of each reply.

Persistent persona

Each companion has a written character that holds under pressure — no prompt leaks, no role hijacking.

Creator controls

A desktop studio with an instant stop button, an OBS overlay, and per-session cost and latency reports.

Where we are

Our first companion is a close-up VTuber for Korean live streaming (CHZZK). The desktop app (Rust + Tauri, Vue) is in active development, with a private test broadcast as the first milestone.