## The problem
Issue deepset-ai/haystack#11040 (closed, 14 comments): **Describe the bug** When running the ResponseChatGenerator in an Agent WITHOUT the exit conditions being a tool (So the tool feeds back into the generator) you get the following. haystack.core.errors.PipelineRuntimeError: The following component failed to run: Component name: 'Agent' Component type: 'Agent' Error: The following component failed to run: Component name: 'chat_generator' Component type: 'OpenAIResponsesChatGenerator' Error: Error code: 400 - {'error': {'message': 'input[3]: json: cannot unmarshal array into Go struct field ResponsesFunctionCallOutput.output of type string', 'typ
## The verified fix
Maintainer anakin87: don't use OpenAIResponsesChatGenerator against non-OpenAI APIs - use the targeted Ollama integration (OllamaChatGenerator) instead, and for reasoning models like Gemma set think=True. A dedicated vLLM integration was in progress and worked with Gemma 4 (Chat Completions API; Responses API support was on the roadmap). Issue closed after the reporter confirmed the integration approach.