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
Known Issues
0.2.0
- Buffered upstream work continues after the client disconnects, so a cancelled request can still incur provider cost.
- Routing-tier attribution is missing from
GET /v1/statsand/metricsfor LLM-classifier judge failures that route to the default target, escalation decisions, andstage_routerfallback decisions. - The retry recovery counter stays at zero after a successful upstream retry.
x-switchyard-session-idis not recorded in native session stats.
0.1.0
- Completed Codex Responses tasks may record
0token usage inGET /v1/stats. - Tool-bearing Codex requests may fail when Switchyard routes them to an upstream that accepts only a fixed set of tool names or schemas.