OPEN LINK ↗
// 1d agoOPENSOURCE RELEASE
NVIDIA Switchyard opens LLM routing
NVIDIA NeMo’s Switchyard is an open-source Rust proxy and library that routes LLM traffic across providers while translating OpenAI and Anthropic APIs. It supports coding agents, A/B testing, operational metrics, and customizable routing algorithms.
// ANALYSIS
Switchyard targets a real pain point as coding agents become multi-model systems: keeping clients stable while providers, models, and routing policies change.
- –Supports Claude Code, Codex, and OpenClaw through a compatible proxy
- –Routes requests using random splits, classifier decisions, or signal-driven stage routing
- –Connects to vLLM, NVIDIA NIM, Ollama, and other OpenAI-compatible endpoints
- –Prometheus metrics expose latency, errors, token usage, and routing overhead
- –Pre-alpha maturity means developers should expect breaking API changes before production readiness
// TAGS
switchyardllminferencedevtoolopen-sourceai-codingapi
DISCOVERED
1d ago
2026-08-12
PUBLISHED
1d ago
2026-08-12
RELEVANCE
9/ 10