OpenAI
Use OpenAI models as execution targets while Vira keeps behavior revision and governance identity separate from provider choice.
INTEGRATIONS
Bring governed behavior to the stack you already run.
An integration should state where Vira executes, what identity it preserves, and what it does not control. Dedicated pages exist only where the repository has enough unique implementation detail.
Anthropic
Compile governed behavior for Anthropic targets without treating provider-specific projection as the canonical policy.
Vercel AI SDK
Resolve governed behavior before generation, preserve native AI SDK streaming, and record the exact behavior identity when generation finishes.
LangGraph
Resolve around graph execution while preserving graph/run correlation and keeping LangGraph as the orchestrator.
OpenAI Agents
Attach exact behavior context at agent lifecycle boundaries without turning Vira into the agent runtime or tool orchestrator.
JavaScript SDK
Resolve behavior, write traces and feedback, and preserve exact revision identity from application code with typed runtime errors.
Expo / React Native
Collect mobile feedback with short-lived application-bound credentials instead of embedding provider or control-plane secrets in the app.
OpenTelemetry
Import runtime observations with explicit provenance and missing-field markers without turning generic telemetry into invented behavior authority.
MCP
Expose reviewed control-plane operations to MCP clients while keeping runtime behavior enforcement and automatic publishing outside the MCP server.