🛡️Data Protection Controls with LangSmith LLM Gateway ✅ Detect, redact, and rep
- Source
- langchain.com
- Author
- LangChain
- Date
Context
LangChain's data-protection controls address two places sensitive information can escape an workflow: requests sent to model providers and the traces of those calls that LangSmith keeps. Its August 10 announcement describes detection and replacement of personal information and secrets before requests reach model providers, plus redaction in the traces the gateway records in LangSmith.
Gateway sits between agents and the models they call, giving a team one place to apply runtime controls across agents, models and providers. According to the access page, Data Protection detects and redacts personal information and secrets, such as , API keys and social security numbers, from model calls, then restores the original values in the model provider's response so that redaction does not break what the agent is doing. The gateway itself is in public beta for all paid LangSmith users, while Data Protection is limited to Enterprise customers who request access through a form. Neither the post nor the page explains how detection works or reports how reliably it catches sensitive data, so these are configurable controls, not a measured guarantee.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- LLM — A large language model — the neural network behind tools like Claude and ChatGPT, trained on huge amounts of text to predict what comes next.
- token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
Checking sign-in…
Loading comments…