Integrations
Connect your existing AI stack.
Nettriva is designed for heterogeneous AI environments. It is built to work with the models, agent frameworks, tool servers, identity providers and security tools you already run — without rebuilding your agents.
Coverage
Built for multi-model, multi-framework environments.
Platforms are grouped by category. Depth of coverage varies by platform and interface; we confirm specifics for your environment during evaluation.
Model Providers
Hosted and self-hosted foundation models
OpenAI
GPT models · Responses API
Anthropic
Claude models · Messages API
Google
Gemini · Vertex AI
Microsoft
Azure OpenAI · Azure AI Foundry
AWS
Amazon Bedrock
Mistral AI
Hosted and open-weight models
Meta Llama
Open-weight models
Self-hosted
vLLM · Ollama · TGI
Agent Frameworks
Libraries and SDKs used to build agents
LangChain
LangChain · LangGraph
LlamaIndex
Agents · Workflows
CrewAI
Crews · Flows
Microsoft
AutoGen · Semantic Kernel
OpenAI
Agents SDK
Google
Agent Development Kit
Protocols & Standards
How agents reach tools, data and each other
Model Context Protocol
MCP servers and clients
Agent2Agent
A2A protocol
OpenAPI
REST tool definitions
OpenTelemetry
GenAI semantic conventions
OAuth 2.0 / OIDC
Delegated access and scopes
Identity
Who agents act for, and with which rights
Okta
Workforce Identity
Microsoft Entra ID
Users · groups · app registrations
Google Workspace
Cloud Identity
SAML · SCIM
Standards-based IdPs
Security Operations
Where findings and audit events are sent
Splunk
Enterprise Security
Microsoft Sentinel
SIEM · SOAR
Elastic
Elastic Security
Datadog
Logs · Cloud SIEM
Collaboration & Ticketing
Where approvals and notifications happen
Slack
Approvals · alerts
Microsoft Teams
Approvals · alerts
Jira
Findings · remediation tasks
ServiceNow
Incidents · change records
Cloud & Runtime
Where agents and their tools run
AWS
IAM · CloudTrail · Lambda
Azure
Entra workload IDs · Activity Log
Google Cloud
IAM · Cloud Audit Logs
Kubernetes
Workloads · service accounts
Vendor, product and platform names are trademarks of their respective owners. They are listed to describe the environments Nettriva is designed to work with. Listing does not imply a partnership, endorsement, certification or official integration with any vendor.
Methods
Observe where possible, enforce where you choose.
Nettriva prefers lightweight instrumentation and the gateways you already run, and adds proxies only where runtime enforcement is needed. All methods start in observe-only mode.
Instrumentation
- Python & TypeScript SDKs
- OpenTelemetry GenAI spans
- Framework callbacks
Gateways & proxies
- Model gateway integration
- MCP proxy
- Egress proxy
Logs & events
- Audit log ingestion
- Webhooks
- Cloud provider logs
APIs
- Identity provider APIs (SCIM)
- SaaS admin APIs
- SIEM & ticketing APIs
Running something not listed here?
Most AI environments include at least one in-house framework, gateway or tool server. Tell us what you run — if it emits traces or logs, exposes an API or can sit behind a proxy, we want to hear about it.
Adopt AI agents without losing control.
See Nettriva investigate a simulated agent incident end to end, and talk with the team building it about your AI environment.