Using the privacy plugin and Seems like every http api request to proxy loads the plugin resources again.
This seems to make the api calls very slow. Can this be loaded into memory for all requests?
2025-09-28 20:31:36,512 - INFO - Received request to /v1/chat/completions
2025-09-28 20:31:36,528 - INFO - Using approach(es) ['privacy'], operation SINGLE, with model gemma3:27b
en_core_web_lg model already downloaded.
2025-09-28 20:31:36,529 - INFO - nlp_engine not provided, creating default.
2025-09-28 20:31:37,915 - INFO - Created NLP engine: spacy. Loaded models: ['en']
2025-09-28 20:31:37,915 - INFO - registry not provided, creating default.
2025-09-28 20:31:37,926 - INFO - Loaded recognizer: CreditCardRecognizer
...
Using the privacy plugin and Seems like every http api request to proxy loads the plugin resources again.
This seems to make the api calls very slow. Can this be loaded into memory for all requests?