2+ day, 12+ hour ago   (591+ words) I want to walk through what LiteLLM does, what an AI security gateway does, where the responsibilities overlap, and how the two layers compose. LiteLLM is an open-source Python package that ships an LLM proxy server. The proxy speaks the…...

3+ day, 12+ hour ago   (509+ words) Benchmarks get compared incorrectly because the metric is left implicit. Name it every time. If a benchmark does not tell you which of these it measured and against what upstream, the number is not comparable to anything. Here is what…...

2+ week, 22+ hour ago   (274+ words) AutoGen ships two logging paths (trace logging for debugging, structured logging for dependable events) plus OpenTelemetry spans for agents and tool calls. None of the three attaches end-user identity to a model call. This piece walks through what AutoGen actually…...

2+ week, 6+ day ago   (554+ words) DeepInspect is a stateless proxy that sits inline between an authenticated user or agent and any LLM. Every HTTP request to a model API, OpenAI, Anthropic, Bedrock, Azure OpenAI, or a self-hosted endpoint, passes through it before the model sees the…...

4+ week, 1+ day ago   (267+ words) IBM watsonx can sit inside a valuable enterprise AI workflow, but DLP requires a decision point that sees the prompt, response, originating identity, and route before content reaches the model. This article defines that enforcement boundary and the evidence security…...