Expand description
llm-dialect: pure sans-I/O translation between LLM wire formats.
Parse an Anthropic (/v1/messages), OpenAI chat (/v1/chat/completions),
or OpenAI Responses (/v1/responses) wire request into the canonical
items model, flatten to the canonical ChatRequest an engine
consumes (dialect::deflate), and render the engine’s responses/
streams back to the caller’s dialect ([dialect::*]).
The mapping is pure: no async runtime, no HTTP types in the public
signatures, no wall clock, no randomness — ids and timestamps are supplied
by the caller at the sanctioned factory boundaries. The optional axum
feature adds the SSE pump (dialect::sse) and per-dialect stream
Response shells for embedders that are an axum service.