Echo voices and Baobab national models — Africa-first language intelligence.
Explore BaobabLugemiAgents
Deploy agents that listen, reason, and reply aloud with cultural context.
Voice Agents are Lugemi’s speaking-agent product: STT → model → TTS turns with Agent Runtime permission allowlists. Built for African languages, accents, and cultural contexts first, with the same /v1 surface for global markets. Builders pick gender, tone/emotion style, and native-accent own:* voices; Emotion Intelligence exposes emotional state, sentiment, and tone on speech or text — honestly as heuristics, not NIST SER.
Play each turn, or play the full agent conversation.
How teams ship with Voice Agents: interactive demos above, step guides below, and the same /v1 surface for APIs and SDKs.
Use case guide
From transcript demo to production speaking agent.
Simulate Voice FAQ turns, then graduate to production with metering and RBAC.
Countries, ethnic groups, and tribes are product metadata — not afterthoughts.
Agent builders select male/female (or neutral) voices, delivery tone/emotion profiles, language, and country/region accent from the own:* registry — Africa-first, with cultural metadata on every pick.
Emotion detect returns emotional state, sentiment polarity, and delivery-tone labels with confidence and honesty notes. Soft audio proxies only — not trained speech-emotion recognition.
