SAN FRANCISCO — xAI is making it dramatically easier to put Grok on the phone. On July 1, the company launched Voice Agent Builder, a no-code platform that lets operators and developers stand up a production-ready voice agent in about two minutes — without writing a single line of code.
The tool runs on Grok Voice and bundles telephony, knowledge retrieval, tools, guardrails and observability into one interface, priced at a flat $0.05 per minute of audio with voices included and no separate platform fee. It is the latest in a rapid cadence of Grok releases that has also brought the assistant to Interactive Brokers for AI-powered market research.
One model, not three stitched together
Most voice stacks chain together three separate services — speech-to-text, a language model and text-to-speech — often from different providers, with each hop adding cost, latency and new points of failure. Voice Agent Builder instead runs on a single speech-to-speech path built for Grok Voice and tightly coupled to the model.
xAI says it trained Grok Voice on the hardest calls it could find: noisy telephony audio, strong accents, interruptions and callers who change their minds mid-sentence, across more than 25 languages. On the company's τ-voice benchmark, Grok Voice Think Fast 1.0 scored 67.3%, well ahead of rival real-time voice models.
Two minutes to a working agent
Setup is deliberately simple. Users write a plain-language description of how calls should flow, then attach documents, tools and guardrails. Knowledge lives in uploadable collections — plain text, Markdown, Word, PowerPoint, Excel and more — that the agent retrieves from during calls.





