The Limerence operator handbook
Install, configure, and run Limerence in your own infrastructure. Architecture, model and database wiring, RBAC, audit, upgrades, and the failure modes you will actually see in production.
By the Limerence team · Living document
Self-host on AWS
Single-region ECS deployment with RDS for state and Secrets Manager for keys.
How it fits together
The four moving parts of a Limerence install and what each one is responsible for.
Connect Postgres safely
Read-only credentials, search-path handling, and re-introspection on schema change.
When things break
The handful of failure modes you actually see in production, and the playbook for each.
Getting started
- 01
Overview & architecture
How Limerence is laid out: control plane, agent runtime, model traffic, your databases. What lives where, and what never leaves your network.
- 02
Self-hosting Limerence
Reference deployments for AWS, GCP, Azure, and on-prem Kubernetes. Sizing guidance, required ports, and the install script.
Connecting
Operating
- 05
Roles, permissions & RBAC
Workspaces, agent visibility, dataset ownership, and the principle of least privilege applied to natural-language access.
- 06
Audit, logging & observability
Every prompt, plan, and SQL execution is recorded. Where to find the logs, how to ship them to your SIEM, and what to alert on.
- 07
Backups, upgrades & disaster recovery
Postgres backup expectations, the upgrade ladder, blue/green for the agent runtime, and recovery time objectives.