fix(ui): expose response cost header and correct prompt caching test copy

Co-Authored-By: Devin AI <158243242+devin-ai-integration[bot]@users.noreply.github.com>
This commit is contained in:
tin 2026-09-17 20:54:01 +00:00
parent 5bec242629
commit 865424fb1c
3 changed files with 9 additions and 1 deletions

View file

@ -205,6 +205,7 @@ LITELLM_UI_ALLOW_HEADERS: Final = [
"x-litellm-guardrail-scan-id",
"x-litellm-guardrail-scan-metadata",
"x-litellm-cache-key",
"x-litellm-response-cost",
]
# Gemini model-specific minimal thinking budget constants

View file

@ -102,6 +102,13 @@ def test_cors_exposes_cache_key_header_to_browser_js():
assert "x-litellm-cache-key" in cors_middleware.kwargs["expose_headers"]
def test_cors_exposes_response_cost_header_to_browser_js():
from fastapi.middleware.cors import CORSMiddleware
cors_middleware = next(m for m in app.user_middleware if m.cls is CORSMiddleware)
assert "x-litellm-response-cost" in cors_middleware.kwargs["expose_headers"]
def test_login_v2_returns_redirect_url_and_sets_cookie(monkeypatch):
mock_login_result = {"user_id": "test-user"}
mock_prisma_client = MagicMock()

View file

@ -91,7 +91,7 @@ const PromptCachingTestCard: React.FC<PromptCachingTestCardProps> = ({
<CardContent>
<CardTitle>Test prompt caching</CardTitle>
<p className="mt-1 break-words text-xs text-muted-foreground">
Sends two identical requests with a ~5k token system prompt and no cache_control of its own. If LiteLLM
Sends two identical requests with a ~9k token system prompt and no cache_control of its own. If LiteLLM
injects the breakpoints, the first call writes the cache and the second reads it.
</p>