You.com Inference Router adds failover and built-in search

Wafer says it is partnering with You.com so that developers can reach Wafer's inference through You.com's new Inference Router. The router is a single API for running open-weight models from several providers. In Wafer's words, access comes through one OpenAI-compatible API, with multi-provider failover and built-in web search.

On its homepage, You.com describes how the router works. A request runs first on You.com's own infrastructure, with backup providers on standby, and if one has trouble the request is rerouted automatically. You.com backs this with a 99.9% uptime SLA. The models it names include Kimi K3, GLM 5.3, DeepSeek V4.1 Flash, Qwen3.8 Flash, Muse Spark 1.3 and Inception Mercury 3. Because the API is OpenAI-compatible, code already written against OpenAI's chat-completions interface can be pointed at it. The built-in search is a server-side web search that a model can call without a separate search integration. You.com also lists 100% US hosting, zero data retention across models and web search, and per-key usage visibility and spend caps.

Wafer's place in this rests on its own post. The You.com page captured here describes the router but does not mention Wafer, and Wafer's post points to a follow-up link that was not captured. The reliability, latency and accuracy claims are You.com's marketing copy, not independent measurements.