Aroovo developer documentation

Aroovo is the multi-tenant, white-label AI agent platform that resellers operate under their own brand. Customers sign in through their own identity provider (Microsoft Entra, Google Workspace, Okta, or Auth0), spend lands on their own LLM account, and bots live in their own Microsoft Teams or Slack workspace. These docs cover everything an engineer needs to integrate the platform, onboard a Brand, and ship agents to a tenant.

The docs are organised top-to-bottom around the four ways developers come at the platform. Quickstart is the first 15 minutes. Concepts explains why the platform is shaped the way it is. API reference documents the surface area you call against. And How-tos covers the specific recipes every Brand admin runs through at least once.

What runs on Aroovo today

The platform ships an in-process agent runtime that calls Anthropic, OpenAI, OpenRouter or OpenClaw on a per-Brand BYO key (see ADR-001). Every Tenant has an isolated audit chain HMAC-signed by a per-deployment secret, an encrypted secrets vault scoped by brandId, a per-tenant rate limit and spend cap, and full BYO bot registration for whichever chat surface the customer uses. Tools are dispatched through the Model Context Protocol — Zapier MCP gives roughly 6,000 SaaS actions through one endpoint and any custom MCP server can be registered alongside.

Where to go from here

Found a gap? The docs live in the public repo at Orion-claw-tech/aroovo-agent-platform — PRs welcome.