Vercel AI Gateway Adds Grok Voice 2.0

// 1h agoPRODUCT UPDATE

Vercel AI Gateway Adds Grok Voice 2.0

Vercel announced the availability of xAI's Grok Voice Think Fast 2.0 on Vercel AI Gateway. The speech-to-speech model reasons concurrently while speaking, enabling low-latency real-time voice applications with unified observability and key management.

// ANALYSIS

Speech-to-speech models that reason while generating audio solve the major latency bottleneck of traditional multi-stage voice pipelines.

–Eliminates the latency penalty of cascading separate Speech-to-Text, LLM reasoning, and Text-to-Speech stages.
–Integration into Vercel AI Gateway gives web developers immediate access to low-latency voice AI without complex backend infrastructure.
–Enables more natural and responsive voice agents for conversational interfaces and automated support workflows.

// TAGS

vercelai-gatewayxaigrokspeechspeech-to-speechllm

DISCOVERED

1h ago

2026-07-30

PUBLISHED

1h ago

2026-07-30

RELEVANCE

8/ 10

AUTHOR

vercel_dev

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

RESEARCH20m ago

TurboVLA hits 32 Hz robotic control on RTX 4090

TurboVLA is a 200M parameter open-source VLA model designed for real-time robotic control at 32 Hz using under 1 GB VRAM on an RTX 4090. By replacing heavy LLM backbones with direct Vision+Language to Action mapping, it achieves a 97.7% success rate on the LIBERO benchmark.

LAUNCH23m ago

Infisical launches Agent Proxy to protect credentials

Infisical launched Infisical Agent Proxy, a credential brokering service that prevents AI agents from directly exposing sensitive credentials. Developers issue dummy API keys to agents, while the proxy intercepts outbound traffic to swap them with real production secrets at the network layer.

UPDATE36m ago

mcp-handler 2.0.0 adds native MCP spec support

Vercel has released version 2.0.0 of mcp-handler, adding native support for the latest Model Context Protocol (MCP) specification. This major update features a stateless fallback mechanism for 2025-era clients and eliminates session storage requirements, enabling easier serverless deployment and improved scalability for MCP tools.