Staff Hub
How the Operio platform is put together, written for the team. Summaries are public; the full notes stay internal.
A grounded comparison of the four agent-interoperability protocols, where each sits in the stack, and what Operio runs in place of each one today — architecture and reasoning, no literal server coordinates.
How Operio watches a fleet of semi-autonomous services from one operator view — a two-tier polling model that separates liveness from usage — architecture and reasoning, no literal server coordinates.
How a fresh client node goes from bare Ubuntu to an isolated, single-ingress GaaS host in one non-interactive run — default-deny firewall, loopback-bound services, and a wildcard TLS reverse proxy — architecture and reasoning, no literal server coordinates.
Architecture walkthrough for Operio Sign — how Documenso CE is deployed, why cloud SMTP requires a VPC relay, and how PDF signing and document storage are wired together.
How a single systemd-managed Node.js process bridges a loopback HTTP API and an external chat channel into one conversational surface — architecture, config patterns, and operational trade-offs.
How Operio runs n8n as a PM2-managed daemon and models a multi-turn intake conversation as an explicit node-graph state machine that proxies the browser to a loopback model gateway and email — architecture and reasoning, no literal server coordinates.
How the Operio v2 stack isolates conversation history, workflow routing, and RBAC in PocketBase collections backed by SQLite — schema, access rules, and the current gap in the vector-store bridge.
How Operio's CrewAI FastAPI service ingests PDF/DOCX/XLSX/PPTX into a local vector store without blocking or OOM-ing a small droplet — background tasks, a concurrency semaphore, and a threadpool for the CPU-bound work. Architecture and reasoning, no literal server coordinates.
How Operio's LiteLLM gateway centralises model access, enforces per-workflow model tiering, and keeps every app oblivious to provider changes — architecture and concepts, no literal server coordinates.
Operio
Operio builds and runs a dedicated AI operations team around the tools you already use.