Solutions
EdTech & Youth Safety Enterprise Security
Site
Company Pricing Contact us Request a demo

Enterprise AI security,
deployed your way

Protect your APIs, govern your AI traffic, and meet strict data residency requirements with a multi-stage detection pipeline that verifies every prompt in under 900ms (p50) and under 1.4s (p95).

Request a demo
98.92%
Attack defense across 1,850 prompts
99.33%
Harmful content classification
<900ms
p50 latency on full pipeline (p95: under 1.4s)
Deployment Models

Zero compromise on data residency

Pick the deployment model that matches your data, compliance, and infrastructure requirements. We forward sanitized traffic, we don't masquerade as your upstream provider.

Managed SaaS

GCP-hosted endpoint

Point your existing LLM API calls at PromptSentry instead of OpenAI, Anthropic, or Google directly. Fast to integrate with our Python SDK. Best for pilots and teams without strict data residency rules.

Customer-hosted K8s

Your cloud, your VPC

Helm chart available for private deployment in your GKE, EKS, or AKS cluster. Prompts and responses never leave your infrastructure. Contact us for access — chart is not yet published to public registries.

Air-gapped

Self-hosted, isolated

Docker container plus local classifier model for isolated deployments. Zero external API calls. Optimized for self-hosted LLM runtimes — contact us for your specific environment.

Universal Python SDK

A few lines of code to wrap any LLM endpoint with our @protect decorator, regardless of which deployment you choose.

pip install promptsentry-sdk
AI Policy Enforcement

Full control over the AI traffic that flows through you

Once your sanctioned AI tools route through PromptSentry, you get prompt-level visibility, policy enforcement, and audit trails on every interaction. Block, redact, warn, or escalate, by user, team, or content category.

  • Prompt-level visibility

    See every prompt your workforce sends to corporate ChatGPT, Claude, or Gemini. Filter by user, time window, or violation type.

  • Policy enforcement, not just detection

    Block, redact, or escalate prompts based on customizable rules. PII detection, secret leakage, and acceptable-use policies, enforced in real time.

policy-engine.log
[INFO] 14:32:01 Prompt received from user@acme.com [INFO] 14:32:01 Running tier-1 fast-path scan... [WARN] 14:32:01 PII pattern detected: SSN [INFO] 14:32:01 Policy lookup: engineering-team [BLOCK] 14:32:01 Action: redact and forward [PASS] 14:32:01 Sanitized prompt forwarded to LLM [INFO] 14:32:01 Audit entry written: scan_99xLk2
mcp-intercept.log
[INFO] 14:32:01 Intercepted tool call: query_customer_db [INFO] 14:32:01 Scanning arguments... [WARN] 14:32:01 Injection pattern in arg: user_id [BLOCK] 14:32:01 Terminated before reaching DB [INFO] 14:32:02 Tool response: fetch_ticket_history [INFO] 14:32:02 Scanning retrieved data for PII... [PASS] 14:32:02 Data sanitized. Forwarding to context.
MCP-Native Defense

Security for the agentic stack

Anthropic's Model Context Protocol connects AI agents directly to your databases, files, and APIs. Every tool an agent can call is a new attack surface. We secure that surface natively.

  • Scan tool call inputs

    Before an MCP tool call reaches your database or API, we scan the call name and arguments. Catches prompt injections hidden in tool parameters.

  • Block indirect prompt injection

    When a tool returns data to the LLM, we scan the response before it enters the model's context. The growing attack vector for agentic systems.

  • PromptSentry as an MCP server

    Install us directly in Claude Code, Cursor, or any agentic framework via stdio. One config file, three tool calls: scan_prompt, scan_tool_call, and scan_response.

Data Privacy

We're a processor, not a controller

Your prompts are processed in memory, evaluated, and discarded. By default we never store the content, only the verdict and metadata you need for audit and compliance.

  • Zero retention by default

    Verdict, latency, confidence, and blocked categories are logged. The prompt content itself is discarded after the scan completes.

  • Per-scan audit IDs

    Every scan generates a unique audit ID. Prove to auditors a prompt was scanned without exposing what it contained.

audit.json
{
  "timestamp": "2026-05-12T14:30:05Z",
  "tenant_id": "org_72bX9aP",
  "scan_id": "scan_99xLk2",
  "verdict": "BLOCKED",
  "categories": ["PII_LEAK"],
  "latency_ms": 4.2,
  "confidence": 0.94,
  "prompt_content": null // zero retention
}
Compliance Posture

Honest about where we are

Enterprise procurement asks tough questions. Here's a straight answer on what we have today and what's on the roadmap.

Live today

  • GDPR compliant by design, data minimization, purpose limitation, right to deletion
  • Data Processing Agreements, signed with every customer
  • Customer-hosted deployments, logs stay in your infrastructure, FNC never sees the data
  • Vendor security reviews, happy to complete CAIQ and similar questionnaires on request

On the roadmap

  • SOC 2 Type II, Q1 2027 target
  • ISO 27001, follows SOC 2
  • BYOK encrypted audit storage, in design for regulated workflows
  • Formal pen test history, Q2 2027 with established firm

See PromptSentry in your stack

30-minute walkthrough with our engineering team. Bring your hardest prompts.

Request a demo