Self-managed Redis Intent Router

Deploy, configure, and operate Redis Intent Router on a self-managed Kubernetes cluster.

Intent Router classifies each incoming user message by intent, so your application can send it to the right agent, workflow, or handler. Developers define intents with example phrases, and their applications send messages to a Router to find the closest match. Before developers can do that, a platform administrator deploys Intent Router and creates Routers and agent keys for them.

This guide is for that administrator. It covers deploying Intent Router on your own Kubernetes cluster with a Helm chart, then configuring, securing, and operating it. To install, start with the prerequisites.

The redis-intent-router Helm chart deploys one tenant's Intent Router control plane and data plane. It deploys no Identity Service: it connects to an Identity Service release that you install separately, which issues and checks agent keys. You provide the Identity Service release, the Redis databases, the embedding provider credentials, and the license.

Applications call the data plane to author intents and match queries, as described in the data-plane API reference. The control-plane API reference documents the administration endpoints for Routers.

These guides use redis-intent-router chart 1.0.0. Its images are tagged with the chart version, 1.0.0.

What you deploy

Use a separate Redis for the control plane's metadata and for the data plane. The data-plane Redis needs the Search and Query engine: use a Redis 8 build that includes it, such as the official redis:8 image, or load the module.

How the components work together

1. Platform admins create Routers through the control plane, then mint agent keys through the Identity Service.
2. The Identity Service asks the control plane to confirm that a Router exists before it mints a key for it.
3. The control plane writes each Router's desired configuration to a mailbox in the metadata Redis. The data plane applies it, builds the Router's vector index, and publishes the result back. The two planes never call each other.
4. Applications call the data plane with an agent key. The data plane checks the key with the Identity Service and reads or writes the Router's data.

A Router serves requests only after the data plane has applied its current configuration. Until then, data requests return 409.

API surfaces

Get started

1. Review the prerequisites.
2. Prepare the configuration.
3. Deploy the chart.
4. Load administrative credentials.
5. Create a Router and an agent key.

Manage the deployment

• Manage Routers through the control plane.
• Rotate or revoke agent keys.
• Check health, rotate credentials, size memory, and upgrade.
• Troubleshoot the deployment.