Data protection & audit for teams shipping AI features

Pass your customer's security review. Ship your AI feature.

Gatekeeper sits in front of your LLM calls and enforces security, privacy, and compliance policies in real time - with audit logs your security team can trust.

Try it for free Get the integration guide

Runs as a proxy. Redaction happens before the provider call.

gatekeeper · sample policy scanExample

Sample scan · outbound prompt

▍

Policy decision

Redact · Review required3 findings

▲ Customer email detected▲ Contract terms detected▲ Financial risk

Recommendation: Redact the flagged context before sending to the external model.

Example output shown.

Live demo

Paste a prompt. Watch Gatekeeper scan it instantly.

See PII detection, jailbreak blocking and policy enforcement running live. Latency is separated into Gatekeeper overhead and the AI provider round-trip.

Prompt

Try something risky: PII, secrets, unsafe instructions, or policy conflicts.

Try an example:

Auto-scan as you type

Scan results

Full pipeline output: jailbreak, IOC, policy and PII checks.

Click an example chip or type a prompt and hit Scan now.

How it works

Add AI policy enforcement in one integration.

Route LLM traffic through Gatekeeper, enforce your rules, and export audit evidence when security asks.

Route traffic

Drop Gatekeeper in front of OpenAI or Anthropic with a one-line change.

Enforce policy

Block, redact, review, or allow prompts based on environment and policy-as-code rules.

Prove compliance

Export audit logs, policy decisions, and evidence for security reviews.

integration.py

# Before - direct, unenforced call
response = requests.post(
  "https://api.openai.com/v1/chat/completions",
  headers={"Authorization": "Bearer $OPENAI_KEY"},
)

# After - routed through Gatekeeper
response = requests.post(
  "https://gatekeeper-production-7dd1.up.railway.app/proxy/openai",
  headers={
    "Authorization": "Bearer $OPENAI_KEY",
    "X-Gatekeeper-Key": "gk_live_xxx",
  },
)

One line of routing. Policy and audit happen in the proxy.

Policy packs

Start with best-practice guardrails. Customize everything.

Policy packs bundle detectors, thresholds, and evidence templates so security and platform teams can move fast without reinventing compliance.

SOC-2

Security + availability posture for AI systems

Access controls for prompts, datasets, and model endpoints
Tamper-evident audit logs for every decision
Change management hooks for policy updates

HIPAA

PHI-aware scanning for healthcare workflows

PII/PHI detection with redaction-first defaults
Minimum-necessary enforcement for tool outputs
Break-glass workflows with reviewer attestations

Why Gatekeeper

Model safety is not your company policy.

Provider filters are generic. Gatekeeper enforces your organization's rules: which data can leave your environment, which tools your agents can call, which environments are allowed, and exactly what evidence gets logged for your security team.

Provider safety filter

Generic, one-size rules
Runs model-side - outside your control
No environment or tool awareness
No audit evidence you own

Gatekeeper policy

Your rules, as code - versioned and reviewable
Enforced in front of the call, before data leaves
Environment-, data-, and tool-aware decisions
Exportable audit logs your security team owns

Integrations

Works seamlessly with your existing tools

Connect Gatekeeper to the platforms security-conscious teams already use - no heavy integration project required.

Workato

Route your Workato AI automations through Gatekeeper - no code required.

Request Workato Connector Access →

Security

Built to be inspected, not just trusted.

Concrete guarantees about how Gatekeeper handles your data - and honest about what isn't certified yet.

Runs as a proxy in front of your LLM calls

Redaction happens before the provider call

Configurable data retention

Audit logs exportable (CSV / JSON)

Policy-as-code - versioned and reviewable

Designed for SOC-2 readiness - SOC-2 Type II in progress

Pricing

Simple pricing. Serious controls.

Upgrade when security and platform teams need enterprise governance - not a prettier dashboard.

Free

For prototypes and local development.

/ month

Live prompt scanning

Community policy packs

Basic audit export (JSON)

Start free

Pro

Popular

For security and platform teams shipping AI to production with guardrails.

$200

/ month

Org-wide policies + environments

SSO (SAML/OIDC) + SCIM

Advanced evidence packs + reporting

Priority support

Start Pro trial

Enterprise

For security-conscious teams in regulated industries and complex deployments.

Custom

Custom detectors + model risk reviews

Dedicated support + SLAs

Security reviews & procurement support

Talk to us

Put a policy gate in front of your AI.

Get the integration guide, then talk to the founder about production enforcement, audit logs, and deployment options.

Get the integration guide Talk to founder