[Feedback] Hallucinated response from Opensourceapp

[techguruagi]

When using the opensourceapp the responses are hallucinated and out of context, while the same sources used for OpenAI returns more or less accurate results. Anyone else is facing the same issue