OpenAI gives its voice agent superpowers to developers - look for more apps soon

Elyse Betters Picaro / ZDNET

Follow ZDNET: Add america arsenic a preferred source on Google.

ZDNET's cardinal takeaways

• OpenAI's Realtime API is present optimized and mostly available.
• You tin effort its latest speech-to-speech model, got-realtime.
• The upgrades amended OpenAI's dependable offerings for developers. 
  

This year, AI agents that tin transportation retired tasks connected behalf of users person been a large focus, with companies perpetually processing offerings that trim the user's workload. To marque these interactions arsenic seamless arsenic possible, galore companies are leaning connected multimodal AI agents, and OpenAI is making processing these products adjacent easier. 

According to the company, OpenAI updated its mostly disposable Realtime API connected Thursday to see much features that let developers and enterprises to physique much reliable dependable agents. Additionally, the institution released its astir precocious speech-to-speech exemplary yet: gpt-realtime. 

The releases: 

RealTime API updates

• What: The upgrades to the Realtime API see enactment for distant MCP servers, representation inputs, and telephone calling done Session Initiation Protocol (SIP), according to the release.
• Why it matters: Ultimately, these expanded capabilities should alteration dependable agents to entree much tools and person much discourse to assistance users. AI tools are lone arsenic adjuvant arsenic the accusation they give, truthful streamlining the process of connecting AI models to information sources is simply a large triumph for developers and users alike. Most importantly, the MCP open-standard ensures that the connections are made, prioritizing idiosyncratic information and privacy. 

A caller speech-to-speech model

• What: OpenAI touted its caller gpt-realtime exemplary arsenic the company's "most advanced, production-ready dependable model." Upgrades see improvements successful intelligence, acquisition following, and relation calling, according to the release. 
• Why it matters: A cardinal tenet of adjuvant dependable assistance and interactions is models that dependable earthy and person the quality to really assistance with tasks. If the caller exemplary works arsenic claimed, it volition alteration a amended acquisition for users.