LangChain, LlamaIndex, Haystack, MCP and Laravel Scout Ship Grounded AI Answers by Default
Every Opensolr AI integration now ships the improved answer behaviour by default. RAG answers come back direct and on topic, without any configuration on your side.
01 What changed in the integrations
The LangChain integration, LlamaIndex, Haystack, the MCP server and Laravel Scout all do two things now, by default:
- A grounded answer instruction is sent by default. The model is told to answer from the retrieved documents, which keeps replies direct rather than discursive.
- Weakly matching documents are dropped before generating. Retrieved results that match poorly no longer reach the model, so answers stay on topic.
This is the default behaviour, not an opt-in flag.
02 Getting it
Update the package you use: pip install -U for the Python integrations, composer update for Laravel Scout.