Error listings
Fix LiteLLM cache_control not forwarding to Anthropic when using Responses API Includes evidence for LiteLLM / Anthropic API troubleshooting demand.
LiteLLM Responses API silently drops cache_control on input_text content blocks during transformation Fix AsyncOpenAI empty api_key error breaking OpenAI-compatible local servers Includes evidence for OpenAI API / Python troubleshooting demand.
AsyncOpenAI(api_key="") raises OpenAIError in v2.34.0 Fix Vercel deploy error 'An unexpected error occurred when running this build' with middleware Includes evidence for Vercel / Next.js troubleshooting demand.
Vercel deploy fails with 'An unexpected error occurred when running this build' with middleware + experimentalServices Fix Claude Code slash commands failing with 'Context limit reached' false positive on 1M context models when context is far below actual limit Includes evidence for Claude Code troubleshooting demand.
Context limit reached Fix LiteLLM Azure OAI authentication error after upgrading to 1.84.0 when using enable_azure_ad_token_refresh Includes evidence for LiteLLM troubleshooting demand.
LiteLLM AuthenticationError: AzureException AuthenticationError - The api_key client option must be set either by passing api_key to the client or by setting the OPENAI_API_KEY environment variable Fix LiteLLM global_max_parallel_requests not enforcing the configured limit, allowing unlimited concurrent requests through Includes evidence for LiteLLM troubleshooting demand.
set global_max_parallel_requests not work Fix Ollama model pull being killed when Chat GUI is closed, and download not resuming from last checkpoint Includes evidence for Ollama troubleshooting demand.
Error: Post "http://127.0.0.1:11434/api/show": dial tcp 127.0.0.1:11434: connectex: No connection could be made because the target machine actively refused it. Diagnose why Claude Code cache expires faster than expected, increasing API costs Includes evidence for Claude Code troubleshooting demand.
Cache TTL silently regressed from 1h to 5m, causing quota and cost inflation Fix Claude Code 'Provider API request failed' error when proxy returns 200 OK Includes evidence for Claude Code troubleshooting demand.
Provider API request failed. (request_id=req_...) Fix Codex CLI hanging forever without generating any response Includes evidence for Codex CLI troubleshooting demand.
Codex CLI hangs indefinitely on all prompts, no response generated Fix Gemini CLI 403 permission denied despite active Google One AI Premium subscription Includes evidence for Gemini CLI troubleshooting demand.
403 PERMISSION_DENIED for Google One AI Premium subscriber Fix Gemini CLI 429 rate limit error on free or Pro accounts Includes evidence for Gemini CLI troubleshooting demand.
429 RESOURCE_EXHAUSTED on Google AI Pro account after full local reset Fix LiteLLM 401 Unauthorized when querying single model info via /v1/models/{model} Includes evidence for LiteLLM troubleshooting demand.
Only proxy admin can be used to generate, delete, update info for new keys/users/teams. Route=/v1/models/{model} Fix LiteLLM A2A /a2a/{id}/message/send returning -32603 Stream consumed error Includes evidence for LiteLLM troubleshooting demand.
-32603 Internal error: Stream consumed Fix LiteLLM budget enforcement blocking free model discovery API access Includes evidence for LiteLLM troubleshooting demand.
Budget enforcement blocks model discovery endpoints (allow accessing free models) Fix LiteLLM ghost models not being cleared when using --num_workers > 1 Includes evidence for LiteLLM troubleshooting demand.
Ghost models with --num_workers > 1: Deleted models not cleared from other workers' local cache Fix LiteLLM not returning Retry-After header when all deployments are in cooldown Includes evidence for LiteLLM troubleshooting demand.
No Retry-After header on RouterRateLimitError (all deployments in cooldown) Handle Bedrock streaming errors gracefully in Anthropic SDK Includes evidence for Anthropic API troubleshooting demand.
ValueError: Bad response code, expected 200: {'status_code': 400, ':exception-type': 'internalServerException'} Fix Claude Code MCP server [object Object] notification error Includes evidence for Claude Code troubleshooting demand.
MCP server error notification renders as "[object Object]" instead of error message Fix Claude Code Opus 4.7 streaming stall that crashes every session Includes evidence for Claude Code troubleshooting demand.
stream_idle_partial lastChunkAgeMs=15001 bytesTotal=661 / sdk_stream_ended_no_result had_error=true