AI Agents Administration Guide
Origin by Corvus provides enterprise-grade AI Agent copilot capabilities (origin.ai).
Overview
AI Agents in Origin by Corvus enable intelligent team assistance, message summarization, semantic search, and automated operational runbooks while preserving complete data sovereignty.
Supported Providers
Origin supports both public and private self-hosted model backends:
- Self-Hosted / Local LLM: vLLM, Ollama, LocalAI (Recommended for air-gapped and sovereign deployments)
- Azure OpenAI: Enterprise-isolated Azure tenants
- AWS Bedrock: IAM-governed foundation model endpoints
- OpenAI-Compatible: Any HTTP API following OpenAI schema standards
Configuration Steps
- In the System Console, navigate to Plugins > AI Copilot / Agents.
- Set Enable Plugin to
true. - Select your Model Backend Provider.
- Enter the API Endpoint URL and API Key / Credentials.
- Configure model parameters (Default Model, Max Context Tokens, Temperature).
- Save settings to activate the agent across channels.
Security & Sovereignty Guardrails
- Zero External Egress: When paired with a self-hosted LLM inside your VPC or tactical perimeter, no telemetry or message content leaves your network.
- Role-Based Access: Restrict agent invocation to specific channels, teams, or enterprise roles.
- Audit Logging: All agent interactions and tool invocations are logged to the central Origin audit stream.