docs: chat_completions extra_body names the skeleton carve-out

chat_completions shares the seam that now refuses the skeleton keys; its docstring still promised an unconditional merged-last win. Same sentence as chat().

Claude-Session: https://claude.ai/code/session_01BAmVWYKoSnFEydbMjuZjHc
This commit is contained in:
Ray
2026-09-03 17:17:19 +08:00
parent c1ad980505
commit f9072f9fff
+4 -1
View File
@@ -1253,7 +1253,10 @@ class PageIndexClient:
OpenAI-compatible backends take them verbatim in the
request body; LiteLLM-routed providers take them as
LiteLLM's own params (mapped or refused per provider).
Credentials belong in ``backend``, never here.
The managed prompt, conversation and tools are not
fields here (``system`` / ``instructions`` / ``input`` /
``messages`` / ``tools`` are refused). Credentials belong
in ``backend``, never here.
extra_headers: Own-model chat only — extra HTTP headers merged into
each backend request; caller headers win. One exception:
LiteLLM's anthropic adapter owns the ``anthropic-beta``