Credentials pasted into copilots
Detect provider keys, cloud access keys, GitHub tokens, JWTs, private keys, and high-entropy secret-like values before they reach an upstream model.
AI data loss prevention
Stop sensitive data before the provider receives it.
Entrovik inspects normalized AI requests and responses in the traffic path, then denies, redacts, mutates, warns, or annotates according to centrally managed policy.
Why teams act
AI adoption creates a new control surface. Put enforcement in the path.
Entrovik gives security, platform, compliance, and application teams a shared technical boundary without coupling enterprise policy to one provider SDK.
PII crossing uncontrolled boundaries
Identify email addresses, phone numbers, US SSNs, Canadian SINs, card numbers, IP addresses, and configured organizational identifiers.
Code and project language leaving the business
Apply source-code, confidential-term, and custom-pattern policies to requests as well as model-generated responses.
Enforcement model
Central policy with application-level context.
Every decision can use normalized content plus tenant, identity, team, application, provider, model, stage, and configured policy metadata.
01Bidirectional enforcement
Run ordered policy stages before provider dispatch and again before content is returned to the caller.
02Explainable decisions
Return policy, action, reason, field, rule, and control reference instead of an opaque 403.
03Content-minimizing evidence
Audit the identity, provider, model, decisions, usage, and outcome without retaining full content by default.
04Policy simulation
Measure would-block and would-redact behavior against labeled datasets before changing production traffic.
In the request path
One decision lifecycle. Every interaction.
Policies remain modular packages. The core resolves, executes, contains, explains, measures, and audits them.
- 01
Normalize text, multimodal parts, and tool-call arguments into a provider-neutral policy input
- 02
Execute the tenant's ordered WASM request pipeline with bounded time, memory, I/O, and concurrency
- 03
Apply denial or content mutation before any provider credential is used
- 04
Inspect the normalized model response through the response pipeline
- 05
Seal content-free decision evidence into the tenant audit chain
Technical FAQ
Questions teams ask before deployment.
Is Entrovik a replacement for endpoint DLP?
No. It is the AI traffic enforcement point. Endpoint, network, SaaS, and data-store controls remain complementary.
Can a policy redact instead of block?
Yes. Each policy can support actions such as warn, redact, mutate, or deny, and its failure behavior is explicit.
Does DLP work on model output?
Yes. Response policies execute before protected content returns to the client.
Can we test our own identifiers?
Yes. Policy configuration supports custom identifiers and patterns, while the WASM ABI allows independently developed detectors.
Put it in front of a real workflow
Turn your AI policy into an enforced decision.
Bring one provider path, one policy requirement, and the architecture your reviewers need to trust.