It should be easier to use Vigil in a more passive manner where all LLM requests/prompts and responses pass through the respective scanners, instead of needing to submit prompts/responses to the Vigil API server. Maybe build it like a drop-in replacement for OpenAI's API schema...
It should be easier to use Vigil in a more passive manner where all LLM requests/prompts and responses pass through the respective scanners, instead of needing to submit prompts/responses to the Vigil API server. Maybe build it like a drop-in replacement for OpenAI's API schema...