Consent-Purpose Binding
- Define purpose catalog aligned to business use-cases
- Bind allowed purposes at project, agent, and tool scope
- Manage consent grants per data principal and purpose
India's First DPDP-Native AI Gateway
Unified AI Control Plane for
Agents Authorization
DPDP-native unified AI control plane offering consent-purpose binding, zero-trust authz, real-time ABAC, RBAC and immutable audit logs.
>pip install oriv-ai
Supports 100+ LLM providers worldwide
OpenAI
Anthropic
Gemini
Ollama
Bedrock
Mistral
Groq
Cohere
Hugging Face
OpenAI
Anthropic
Gemini
Ollama
Bedrock
Mistral
Groq
Cohere
Hugging Face
Compliance
DPDP compliance simplified
ORIV embeds regulatory compliance into the AI request lifecycle so your teams can innovate freely while staying within guardrails.
Runtime Policy Enforcement
- Enforce granular ABAC and RBAC based policies on every request
- Block when identity or active consent is missing
- Apply most-restrictive policy outcome across scopes
Audit Ready Infrastructure
- Immutable, tamper-proof, audit-ready logs
- Request-level trace context with consent/policy decisions
- Queryable logs for audits and regulatory response
Policy Engine
Policy Playground
Define policies with intuitive forms. Configure guardrails, budget limits and routing.
policy-builder
PII Masking
Auto-detect and mask sensitive data
$
Budget Limit
Monthly cap$100
Fallback Model
claude-3-haiku
#
Max Tokens
4096
Rate Limit
120 requests/min
Prompt Injection Guard
Block malicious tool-use prompts
Simulate and analyze behavioral changes over a 30-day period using historical request data
policy-simulation
Replay windowLast 30 days
Total Replayed
24,580
12% from last month
Would Allow
21,124
90.1% of total traffic
Would Block
2,318
9.9% high risk
Decision Distribution
Simulation output
Allow 90.1%Block 9.9%
Top Blocked Policies
Count
PII Masking1,042
Consent-Purpose Binding786
Tool Access Restrictions490
One click Policy Enforcement
Security
Everything you need to secure AI at scale
Built for security teams that need enforceable controls, clear accountability, and production-grade protection across every AI interaction.
Role-Based Access Control
Enforce least-privilege access with project-scoped roles for users, service keys, and runtime agents.
Attribute-Based Access Control
Evaluate real-time context and request attributes to allow or deny tool execution.
Memory-Based Attribute Control
Apply dynamic access checks using session memory attributes so tool permissions adapt to runtime context.
Default-Deny Enforcement
If policy checks fail or are unavailable, access is blocked automatically.
Credential Isolation from LLMs
The LLM cannot read, retrieve, or return tool credentials at any stage.
Forensic Audit Visibility
Each interaction is logged with request, policy, and guardrail decisions for investigations.
Ready to ship production AI?
Zero-config onboarding. Drop-in compatibility. Cloud-agnostic deployment. Start securing and optimizing your AI applications in minutes.