Repo · Safety & guardrails

jev-model-tokengate

Streaming LLM proxy that holds tokens in a sliding buffer, runs semantic safety criteria (TypeSafe Jev by default) on each window, and releases or…

Open github.com ↗

How builders describe it

An OpenAI-compatible proxy that sits between your LLM and your users. It evaluates each sliding window of tokens while the response is still streaming and cuts the stream before a violating token can reach the screen.

Streaming LLM proxy that holds tokens in a sliding buffer, runs semantic safety criteria (TypeSafe Jev by default) on each window, and releases or aborts before violating tokens reach the client—demonstrating inline interception versus post-hoc redaction.

The decision Jev makes

Allow, flag or escalate a message, tool call or action.

Where it fits

Screening messages, tool calls and outputs before they cause trouble. Jev gives a fast, calibrated verdict that can sit in front of every request or agent action. All 131 safety & guardrails projects →

Related projects