Local Deployment

The fastest way to run Routero AI on your own machine — for evaluating the platform, local development, or air-gapped environments that need the full control plane without a cloud dependency.


When to choose this

  • Evaluation — explore Routero’s routing, policy, and Advanced Features before committing to a cloud deployment.
  • Local development — run the gateway alongside your application during development so your dev environment matches production.
  • Air-gapped / offline — environments with no outbound internet access. Models are served from local endpoints (e.g. Ollama); the gateway enforces the same policies as cloud.
  • CI/CD integration testing — spin up the gateway in a pipeline to integration-test your application against a real Routero instance.

Local Deployment is not intended for production traffic. For production, use Routero Cloud, Single-Tenant Cloud, or Private Deployments.


What you need

Component Requirement
Postgres v14+ with pgvector (for audit log and key storage)
Redis v7+ (rate limiting and caching)
Routero proxy Container image — provided in the deployment package
Admin key A MASTER_KEY you choose; used to authenticate management API calls

No cloud account, no Terraform, no external services required. All traffic stays on your machine.


Capabilities in local mode

All Routero features work locally with no configuration changes required:

Feature Works locally
Routing, failover, load balancing
Policy routing (YAML rules)
Budgets and spend tracking
Virtual API keys and orgs
Guardrails (including Presidio PII)
Prompt Management
Token Saving (compression + caching)
Memory-as-a-Service (Mem0 / Cognee) ✓ with pgvector
Admin console ✓ served locally at /_experimental/out/

Local model endpoints (Ollama, LM Studio, vLLM, any OpenAI-compatible server) are supported as first-class providers — add them under Models → Provider Keys in the admin console.


Getting the deployment package

Contact solutions@routero.ai to receive:

  • Container image access (private registry)
  • docker-compose.local.yml quick-start for getting Postgres + Redis + the proxy running in minutes
  • License key for local use

Reference Architecture for the full component topology · Advanced Features to explore what’s available once running