Skip to main content
We’re building a standard, not just a tool. The Agent Identity Protocol aims to become the universal identity layer for the Internet of Agents (IoA).

​
Vision

Establish AIP as the de facto standard for agent identity, authentication, and authorization—submitted to the IETF for standardization and adopted across the AI ecosystem.

​
Release Timeline

​
v0.1: Localhost Proxy ✅

Status: Released (2026-01-20) Goal: The “Little Snitch” for AI Agents Provide local policy enforcement for individual developers securing their AI agents.
  • ✅ Tool allowlist enforcement
  • ✅ Argument validation with regex
  • ✅ Human-in-the-Loop (macOS, Linux)
  • ✅ DLP output scanning
  • ✅ JSONL audit logging
  • ✅ Monitor mode (dry-run testing)
  • ✅ Cross-platform support (macOS, Linux, Windows)
  • ✅ Cursor/Claude Desktop integration
  • ✅ Example policy library
Use Case: Individual developers securing local AI agents during development.

​
v0.2: Kubernetes Sidecar 🚧

Status: In Progress Goal: The “Istio” for AI Agents Enable enterprise deployment with orchestration and observability.

Kubernetes Integration

  • Helm chart for easy deployment
  • NetworkPolicy integration
  • Pod security policies
  • Init container pattern

Observability

  • Prometheus metrics export
  • OpenTelemetry tracing
  • Grafana dashboard templates
  • Alert rule examples

Multi-Tenancy

  • Namespace-scoped policies
  • RBAC integration
  • Tenant isolation
  • Policy inheritance

High Availability

  • Stateless proxy design
  • Horizontal pod autoscaling
  • Health checks and readiness probes
  • Graceful shutdown
Use Case: Platform teams deploying agent infrastructure at scale. Target Release: Q2 2026

​
v1.0: OIDC / SPIFFE Federation 📋

Status: Planned Goal: Enterprise Identity and Federation Integrate with existing identity infrastructure for production-grade deployments.
1

Workload Identity Federation

  • SPIFFE/SPIRE integration for workload identities
  • Kubernetes ServiceAccount binding
  • AWS IAM Roles for Service Accounts (IRSA)
  • GCP Workload Identity
  • Azure Managed Identity
2

Centralized Policy Management

  • Policy registry with versioning
  • GitOps integration (ArgoCD, Flux)
  • Policy-as-Code validation
  • Automated policy rollout
3

Multi-Tenant Audit Aggregation

  • Centralized audit log aggregation
  • SIEM integration (Splunk, Datadog, Elastic)
  • Compliance reporting (SOC 2, GDPR, HIPAA)
  • Forensic query capabilities
4

Federation Support

  • Cross-organization agent identity
  • Trust framework for agent delegation
  • Federated policy enforcement
  • Inter-domain audit trails
Use Case: Enterprises requiring compliance, federation, and centralized governance. Target Release: Q4 2026

​
Future Extensions

These features are being considered for post-v1.0 releases based on community feedback.

​
Network Egress Control

Problem: MCP servers can currently make arbitrary network calls. Solution:
  • Declarative egress allow/block lists
  • DNS-based filtering
  • Container network isolation
  • Circuit breaker patterns
Target: v1beta1

​
Agent Capability Attestation

Problem: No cryptographic proof of agent capabilities. Solution:
  • Hardware-backed attestation (TPM, SGX)
  • Manifest signing and verification
  • Runtime integrity monitoring
  • Capability certificates
Target: v1.1

​
Multi-Agent Workflows

Problem: Complex workflows involving multiple agents lack coordination. Solution:
  • Workflow-scoped policies
  • Agent-to-agent delegation
  • Composite audit trails
  • Workflow orchestration primitives
Target: v1.2

​
AI Model Governance

Problem: No visibility into which models are used by agents. Solution:
  • Model provenance tracking
  • Model allowlists by vendor/version
  • Inference cost tracking
  • Model bias auditing hooks
Target: v2.0

​
IETF Standardization Track

Our goal is to submit AIP to the IETF for standardization:
1

Draft Specification

Complete formal specification documentStatus: v1alpha2 complete
2

Reference Implementations

Multiple independent implementationsStatus: Go implementation stable, Rust planned
3

Conformance Testing

Comprehensive test suite for interoperabilityStatus: In progress
4

Community Adoption

Real-world deployments and feedbackStatus: Early adopters testing
5

Internet-Draft Submission

Submit to IETF for reviewStatus: Planned for Q3 2026

​
How You Can Help

We welcome contributions at all levels:

Try It

Deploy AIP in your environment and provide feedback on usability and edge cases.

Contribute Code

Implement features from the roadmap or build integrations with your favorite tools.

Spread the Word

Share your experience, write blog posts, or speak at conferences about AIP.
See our Contributing Guide to get started.

​
Stay Updated

Follow development progress:
Roadmap items and timelines are subject to change based on community feedback and priorities.

Build docs developers (and LLMs) love

Get started for freeTalk to us