a11oy · WALLPA — the Voice · Doctrine v13 §2.2

WALLPA — the Voice

Quechua wallpay = to create / to invent; wallpa = the created/expressed thing. The expression organ: it renders the argmax-selected action into one shared voice, with a fidelity gate. Factor: Wallpa(a)=𝟙[render(a)⊑sanctioned(a)]·(1−drift_out) ∈ [0,1].

★ Demo — 30s narration of the PURIQ master formula

POST /api/a11oy/wallpa/narrate-doctrine (Hatun-Willay narrator)
— renders Doctrine v13's PURIQ master formula as audio in the Hatun-Willay synthetic timbre —

Per-organ voice profiles (synthetic timbres)

GET /api/a11oy/wallpa/voices
— one synthetic timbre per organ: Amaru / Yuyay / Killinchu / Hatun-Willay / Chaski / Wasi-Rikuq —

Speak — synthesize any text

POST /api/a11oy/wallpa/speak · SSE: GET /speak/stream
— synthesize speech; audio + transcript hashed into a Khipu receipt —
Honest labels (Doctrine v13 §5 + HARD RULES): OPEN-SOURCE TTS ONLY (Piper / Coqui XTTS-v2 / OpenVoice; deterministic synthetic-timbre fallback when no model is installed, so the endpoint is never vapor). All per-organ voices are SYNTHETIC TIMBRES — never clones of real people, no proprietary voice cloning without consent. Every utterance emits a Khipu receipt = SHA3-256 of (audio ‖ transcript); signature is DSSE PLACEHOLDER.