Portfolio · Voice AI Stacks
Voice AI Stacks ready to use
Each stack below is a working demo, built by Raphael Bruno. Pick the combination that fits your use case.
Live Demo · Available Now
Hume EVI 4-mini
Native pt-PT voice, adaptive prosody, end-to-end.
Hume EVI 4-mini · Claude Sonnet 4.5
Try it →Test · Gemini Live
Gemini Live via LiveKit
Native audio model, multilingual, phone-line candidate.
Gemini 2.0 Flash · LiveKit · pt-PT
Try it →Test · ElevenLabs ConvAI
ElevenLabs Conversational AI
STT+LLM+TTS pipeline, pt-PT voice, ElevenLabs ecosystem.
ElevenLabs ConvAI · pt-PT voice
Try it →Pipeline · Vapi
Vapi
Browser voice orchestrator — Claude thinking, ElevenLabs Portuguese voice speaking.
Vapi · Claude Sonnet 4 · ElevenLabs
Try it →Pipeline · Retell
Retell AI
Another browser pipeline orchestrator, to compare flow and latency.
Retell · Gemini · ElevenLabs
Try it →Telephony · Twilio
Twilio
Native ConversationRelay — the agent answers and dials over a real phone line.
Twilio ConversationRelay · Claude · ElevenLabs
Try it →