Inworld AI: Ukubuyekezwa ku-Quasa — Izwi Lesikhathi Sangempela Nabalingiswa be-AI

#Quasa #QUA #inworld

I-Inworld AI iyinkundla ye-AI esebenza ngesikhathi sangempela egxile ekuxhumaneni ngezwi okufana nokwabantu nakubalingiswa abahlakaniphile. Yasungulwa ngo-2021, futhi ihlinzeka abathuthukisi ngobuchwepheshe obusezingeni eliphezulu bokuguqula umbhalo ube yinkulumo, ukuguqula inkulumo ibe umbhalo, ukuqondisa izicelo kuma-LLM, kanye ne-Character Engine eyenzelwe ama-NPCs aguquguqukayo, ama-ejenti ezwi, nokuhlangenwe nakho okusebenzisanayo.

Imikhiqizo namasevisi ayinhloko kuhlanganisa i-Realtime TTS-2 (inkulumo yekhwalithi ephezulu enokubambezeleka okuphansi, eqondiswa ngolimi lwemvelo futhi esekela izilimi ezingu-200+), ukukopisha izwi, i-Realtime STT ehlaziya imizwa nomongo, i-LLM Router ehlakaniphile, kanye ne-Realtime API ehlanganisiwe yezingxoxo zezwi nezwi lapho izinhlangothi zombili zingakhuluma futhi zilalele ngasikhathi sinye. I-Character Engine yenza kube nokwenzeka ukuba nabalingiswa be-AI nama-NPCs anenkumbulo ehlala ikhona, ubuntu, imizwa nemigomo — okuyenza ifaneleke kakhulu imidlalo eyakhiwe nge-Unity nange-Unreal Engine. Ama-SDKs, amathuluzi abonakalayo nezifanekiso kusiza amaqembu akhiphe imikhiqizo ye-AI esebenzisanayo ngokushesha.

I-Inworld isetshenziswa izitudiyo zemidlalo, izinhlelo zokusebenza zabathengi, amathuluzi okufunda izilimi nezinkundla zokuzijabulisa ezidinga izwi eliveza imizwa elinokubambezeleka okuphansi kanye nabalingiswa ngezinga elikhulu. Inkampani igcizelela amanani ancintisanayo, ukuvikeleka kwamabhizinisi (izinketho ze-SOC 2, GDPR, HIPAA), kanye nokusebenza okulungele ukusetshenziswa emikhiqizweni esisebenzayo.

I-Inworld AI ifaneleka kakhulu abathuthukisi bemidlalo, abadali bemidiya esebenzisanayo, abakhi bama-ejenti ezwi, amaqembu athuthukisa izinhlelo zokusebenza zabathengi, nanoma ubani odala ukuhlangenwe nakho kwezingxoxo zesikhathi sangempela noma okuqhutshwa abalingiswa.

Amaphuzu avelele

• I-TTS yesikhathi sangempela ekleliswe ku-#1: ukubambezeleka okungaphansi kuka-200 ms nokusekelwa okuqinile kwezilimi eziningi.
• Character Engine: ama-NPCs aqhubeka ekhona futhi anemizwa, nama-ejenti e-AI ahlakaniphile.
• I-Realtime API ehlanganisiwe: uchungechunge lokucubungula lwe-STT + LLM + TTS oluhlanganisiwe lwezingxoxo ezigeleza ngokushelela.
• Uhlelo olubanzi lwamathuluzi namasevisi: ama-SDKs e-Unity, Unreal, ne-Web; ukukopisha izwi nokuqondisa okuthuthukile.
• Isisekelo esikwazi ukwandiswa: amanani ancintisanayo asekelwe ekusetshenzisweni nengqalasizinda elungele ukusetshenziswa emikhiqizweni esisebenzayo.

Okungase kudinge ukucatshangelwa

• Umsebenzi wokuklama: imiphumela emihle kakhulu idinga ukuklanywa kwabalingiswa okucatshangelwe kahle nokuhlanganiswa kobuchwepheshe.
• Izidingo zamazinga: ukusetshenziswa okuthuthukile emikhiqizweni esisebenzayo (ukusebenza okuningi ngasikhathi sinye, amazwi enziwe ngokwezifiso) kutholakala emazingeni akhokhelwayo.
• Umkhakha okhethekile: igxile ezwini lesikhathi sangempela nakubalingiswa kunasekukhiqizeni okuhloselwe imisebenzi ehlukahlukene ngaphandle kokuxhumeka ku-inthanethi.

Isinqumo sisonke: izinkanyezi ezingu-4.7/5

I-Inworld AI ingenye yezinkundla ezinamandla kakhulu zezwi nabalingiswa besikhathi sangempela ngo-2026. Ngokuhlanganisa inkulumo yekhwalithi ephezulu enokubambezeleka okuphansi ne-Character Engine eyakhelwe le njongo, isiza abathuthukisi badale ukuhlangenwe nakho okusebenzisanayo okuheha kakhulu futhi okufana nokwabantu emidlalweni nasezinhlelweni zokusebenza zabathengi.

Qala: https://quasa.io/projects/inworld-ai