feat(llm): route adapters by provider

docs/architecture.md

@@ -125,7 +125,7 @@ Durability is a plugin concern. Persistence backends buffer synchronous `session
 
 Messages are arrays of typed content blocks (`text`, `reasoning`, `tool-call`, `tool-result`). The union derives from the merge-extensible `ContentBlockMap`; the same pattern types `MessageSource`, `FinishReason`, `TurnTrigger`, and `TurnEndReason`. New block types are coordinated across adapters, UI bridges, compaction pricing, and persistence, so block types remain a repo-wide contract.
 
-Streaming is a raw chunk protocol (`block-start` through `finish`) with `BlockAssembler` as the shared chunk-to-block assembler. The loop logs raw chunks while assembling them for dispatch. `LlmAdapter` is the provider seam: subclass, implement `stream()`, and register with `ctx.llm.registerAdapter(models, adapter)`. StreamChunk conventions live in [llm-streaming.md](core-data-structures/llm-streaming.md).
+Streaming uses raw chunks (`block-start` through `finish`) and `BlockAssembler`. The loop logs and assembles chunks, storing assistant provider/model provenance plus replay state. An `LlmAdapter` implements `stream()` and registers routes with `ctx.llm.registerAdapter(providers, adapter)`; requests route by `provider`, while the adapter resolves `model`. Replay state reaches a target only when both routes map to the same adapter instance, which owns validation and conversion. The contract lives in [llm-streaming.md](core-data-structures/llm-streaming.md).
 
 ## Extension And Composition