Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimization.
Rust
Commits on main
feat(server): advisor_gate stats projection and e2e coverage (#383)
1700f62 eric-liu-nvidia committed 3 months ago
feat(libsy): advisor_gate review-gate algorithm (#371)
4dd5245 eric-liu-nvidia committed 3 months ago
feat(server): add Dockerfile for switchyard-server container image (#421)
61ee1b3 Peter Pan committed 3 months ago
fix: normalize empty content-filter responses instead of raising (#358)
fa006f2 Andrew White committed 3 months ago
feat(libsy): record task_kind and agent_role on the run span (#249)
a3ac8b3 yanyanz89 committed 3 months ago
fix(translation): accept SSE data fields with no space after the colon (#447)
1149e8e monem committed 3 months ago
fix(client): strip api-key and OpenAI org/project headers before forwarding (#420)
70fe89c Atharva-Kanherkar committed 3 months ago
fix(llm-client): detect native sglang context-overflow messages (#426)
2d0d4b2 Francesco Parisio committed 3 months ago
refactor(protocol): use typed HTTP status codes (#457)
05533a5 Bhuvan Agrawal committed 3 months ago
fix(translation): preserve chat reasoning details (#415)
f9b2df6 Alex Steiner committed 3 months ago
docs(changelog): note packaging extras removal in Unreleased (#433)
f735d9d LeonSGP committed 3 months ago
feat(libsy-llm-client): Move retry logic from libsy to libsy-llm-client (#431)
9ad6744 Graham King committed 3 months ago
fix(server): use normalized session IDs in routing stats (#430)
f1f9e04 Alex Fournier committed 3 months ago
feat(benchmark): pre-bake hermes agent into dataset images for closed-book runs (#350)
0951d0a GDRS committed 3 months ago
feat(server): forward inbound Anthropic auth (#372)
e1eb3d1 Elyas Mehtabuddin committed 3 months ago
fix(protocol): normalize nested metadata strings (#422)
c421dfe Ngo Quoc Viet committed 3 months ago
Revert "ci: base full CI on changed paths (#403)" (#414)
a12a741 nachiketb-nvidia committed 3 months ago
fix(translation): preserve multimodal tool results (#389)
c689466 Elias Shieh committed 3 months ago
feat(examples): update LiteLLM for run stream (#412)
e9767d6 nachiketb-nvidia committed 3 months ago
fix(server): reject authentication headers in extra_headers (#360)
28ef3a0 Elyas Mehtabuddin committed 3 months ago
fix: Stamp the selected ModelId on LlmRequest much earlier (#393)
224287b Graham King committed 3 months ago