Self-host

One binary.
Your infrastructure.

Chimes targets three deployment shapes from a single configuration. The same engine runs managed in the cloud, on your own Postgres, or fully local — your call, not ours.

Read the guide Talk to us

Cloud

Managed on the L1fe platform — identity, IAM, and billing handled. Set mode = "cloud" and go.

Self-hosted

Your own Postgres, persistent across restarts and processes. Set mode = "postgres" with your secrets.

Local / dev

In-memory, perfect for sandboxes and CI. Set mode = "local" — never allowed in production.

Why self-host

The control you actually want.

For teams with compliance needs, data-residency rules, or simply a preference to own their stack — self-hosting is a first-class path, not an afterthought.

One binary

The entire engine ships as a single Rust binary. No sprawling microservice mesh to operate.

Your infrastructure

Run it in your VPC, on your Postgres, with your KMS and identity providers behind their adapters.

Same code path

Cloud, self-hosted, and local run the identical engine. What you test is what you ship.

Open-source core

Layer 1 is Apache-2.0. Read it, audit it, extend it, and own your deployment.

Up in minutes

Three commands to a running engine.

Start the dependencies, run the server, open the dashboard. Add your model key when you're ready to turn on AI resolution.

getting-started

$ docker compose up -d postgres nats otel-collector
$ cargo run -p chimes-server          // binds :8080
$ cd apps/dashboard && npm run dev    // :3000

[adapters.iam]
provider = "keystone"
mode     = "postgres"                 // durable, self-hosted

“It runs entirely in our VPC, on our Postgres, with our keys. Compliance signed off in a week.”

Head of Platform · Regulated SaaS

Questions

The short answers.

A single binary — chimes-server — plus PostgreSQL and NATS. Bring it up with docker compose for dependencies and cargo run (or the released binary) for the engine. The dashboard is a standard Next.js app.