Enterprise AI Gateway
with Production-Grade Governance
The only AI gateway built for governance-first enterprises. Model allowlisting, guardrails for PII/credentials, user spending limits, RBAC, audit logging, and SOC 2 Type II complianceâproduction-ready in minutes, not months.
Trusted by teams at
What is an Enterprise AI Gateway?
An Enterprise AI Gateway is a centralized governance layer that controls all AI model access across your organization. Unlike generic API gateways or building in-house, Requesty provides:
Model Governance
Approved model allowlisting by provider, location, capabilities, and cost. Users see only approved models.
Guardrails & Safety
Automatic PII/credential detection and blocking in inputs and outputs. GDPR, PCI DSS, HIPAA compliance.
User & Group Controls
Per-user spending limits ($100 for Sarah, $50 for Jake), RBAC roles, group-based access, audit logging.
Built vs Buy
Production-ready governance in 5 minutes vs 6+ months of in-house development, ongoing maintenance, and security patching.
AI Governance Framework Alignment
Requesty maps to recognized governance standards that CISOs, Legal, and Risk teams rely on
NIST AI RMF
Govern (GV), Map (MP), Measure (MS) functions through model allowlisting, guardrails, and audit logging
ISO/IEC 42001
AI management system with documented policies, risk assessments, and continuous monitoring
EU AI Act
High-risk AI system controls with human oversight, transparency, and data quality requirements
SOC 2 Type II
Security, availability, and confidentiality controls audited annually by independent third party
What Makes Requesty Different
Purpose-built for AI governance, not retrofitted from API gateways
Governance-First Architecture
Every feature designed for compliance and control. Model allowlisting, guardrails, audit logging, and RBAC are coreânot add-ons.
User-Based, Not Key-Based
Control spending per user ($100 for Sarah, $50 for Jake), not per API key. Track who uses what models, when, and how much they spend.
Real-Time Guardrails
Detect and block PII, credentials, and financial data in both inputs and outputs. Works across all models and providers automatically.
5 Minutes to Production
No months of development, no ongoing maintenance. Change your base URL to router.requesty.ai/v1 and get enterprise governance immediately.
User-Based Spending Control
Set individual spending limits per user, not per API key. Finally get granular control over your team AI spend with role-based access and permissions.
Key Benefits
Individual Budgets: Set $100/month for Sarah, $50 for Jake - per user, not per key
Role-Based Permissions: Admins see everything, developers see only their data
Real-Time Tracking: Monitor spending as it happens, prevent budget overruns automatically
User Management Dashboard
User spending and permissions overview
3
$1,350
$1,090
Sarah Chen
sarah@company.com
Jake Rodriguez
jake@company.com
Emily Watson
emily@company.com
SSO Configuration
Enterprise identity management
Provider
Okta
12 users
2 minutes ago
Azure AD
Available Soon
Coming Soon
12
+3 this week
99.9%
Last 30 days
Recent SSO Activity
User login
sarah@company.com
2 min ago
User provisioned
jake@company.com
15 min ago
Enterprise SSO Integration
Seamlessly integrate with your existing identity provider. Support for SAML, OAuth, and custom authentication flows with automatic user provisioning.
Security Features
SAML & OAuth Support: Works with Okta, Azure AD, Google Workspace, and more
Auto User Provisioning: Automatically create and manage user accounts from your IdP
Audit Logging: Complete audit trail of all authentication and authorization events
AI Model Governance
Control which AI models your team can access. Set policies per team, department, or individual users. Ensure compliance and cost control at scale.
Governance Controls
Model Allowlisting: Specify exactly which models each team can use
Cost Controls: Prevent expensive model usage with automatic limits
Compliance Policies: Enforce data residency and regulatory requirements
Model Access Control
Approved models and compliance overview
3
Allowed
1
Restricted
1
Pending
1
Blocked
OpenAI
4 models ⢠SOC 2
Anthropic
3 models ⢠SOC 2, HIPAA
3 models ⢠SOC 2
Meta
2 models ⢠Limited
Cohere
2 models ⢠Under Review
Mistral
2 models ⢠Non-compliant
Role-Based Analytics & Insights
Powerful analytics tailored to each role. Executives see spend trends, managers see team performance, developers see their usage - all in real-time.
Analytics Features
Custom Dashboards: Each role sees relevant metrics automatically
Usage Trends: Identify patterns and optimize AI spend across teams
Export & Reports: Generate detailed reports for compliance and budgeting
Try Different Views
Switch between admin and user views to see how role-based access controls work in practice.
Analytics Dashboard
Full team visibility and controls
Team Spending Breakdown
Team Members
Sarah Chen
17,500 requests this month
$1200/$1000
Jake Rodriguez
2,800 requests this month
$72/$100
Emily Watson
8,900 requests this month
$280/$250
Alex Kim
1,200 requests this month
$35/$50
Frequently Asked Questions
Common questions about enterprise AI governance, controls, and features
What is an Enterprise AI Gateway?
An Enterprise AI Gateway is a centralized platform that manages all AI model access across your organization. Requesty provides user-based spending limits, model governance (approved model allowlisting), guardrails for PII/credential detection, RBAC with admin and standard user roles, group-based access control, and audit logging. Unlike building in-house, you get production-ready governance in minutes, not months.
How does user-based spending control work?
Set monthly spending limits per user (e.g., $100 for Sarah, $50 for Jake), not per API key. When a user approaches their limit, they receive automatic alerts. Once the limit is reached, their API calls are blocked until the next month or until an admin increases their budget. Admins can view all users' spending in real-time and set default org-wide limits.
What is model governance and approved models?
Model governance lets admins create a curated whitelist of AI models for the organization. Users see only approved models by default and cannot use non-approved models. Admins can approve models based on provider, location, capabilities, pricing, and context window. This ensures compliance, cost control, and strategic model usage across teams.
How do guardrails protect our data?
Guardrails automatically detect and block sensitive information in both inputs and outputs: PII (SSNs, emails, phone numbers), credentials (API keys, database passwords, tokens), and financial data (credit cards, bank accounts). This ensures GDPR, PCI DSS, and SOC 2 compliance. Guardrails work across all API keys and models with real-time scanning and automatic masking.
What are the available RBAC roles?
Requesty currently has two roles: Standard Users (access only personal data, manage own API keys, view personal logs) and Administrators (full org-wide access, manage all users/groups/API keys, configure enterprise features). Future roles will include Group Administrators, Read-Only Analysts, API Key Managers, and Billing Administrators.
How do groups work for access control?
Groups organize users into logical units (departments, projects, teams). Users can belong to multiple groups. Associate API keys with specific groups to control model access and track collective spending. Groups support flexible membership and reflect your company's organizational structure.
Does Requesty support SAML SSO?
Yes. Requesty integrates with Okta (currently live), Azure AD (coming soon), and other SAML/OAuth providers. SSO includes automatic user provisioning, audit logging, and seamless integration with your existing identity provider. 99.9% success rate for authentication in the last 30 days.
What compliance standards does Requesty meet?
Requesty is SOC 2 Type II certified, GDPR compliant, HIPAA-ready, and supports data residency requirements. Our guardrails enforce PII detection, credential masking, and PCI DSS compliance. All authentication and authorization events are logged for audit trails.
Can I set different spending limits for different teams?
Yes. Use groups to organize teams, then set per-user limits within each group. For example, Engineering group members get $200/month, Marketing gets $50/month. Admins can also apply default spending limits org-wide and override them per user or group.
How does Requesty compare to building an AI gateway in-house?
Building in-house requires months of development for user management, SSO integration, RBAC, guardrails, model governance, and audit logging. Requesty provides all these features production-ready in minutes. You avoid ongoing maintenance, security patching, and scaling challenges while getting enterprise-grade governance immediately.
What happens when a user exceeds their spending limit?
When a user approaches their limit (e.g., 80%), they receive an automatic alert. Once the limit is reached, their API calls are blocked with a clear error message. The user's access is restored automatically at the start of the next month, or an admin can manually increase their budget at any time.
Can users see other users' data and spending?
No. Standard users can only see their own data, API keys, logs, and spending. Only administrators have org-wide visibility. This ensures user privacy, data isolation, and compliance with the principle of least privilege. All role-based access is enforced at the platform level.