Building backend services in any language that fits in a container

Most "agentic" platforms make a quiet assumption: build in our framework, our language, our way, or you're on your own.

Ardor's agentic backend takes a different position. The rule, straight from the docs:

Build backend services in any language that fits in a container.

That's the whole constraint. Python, Go, Rust, whatever your stack already runs on — if it fits in a container, Cerebrum (Ardor's copilot) can build it, deploy it, and keep it in the same working context as your other services, domains, and deployments.

"Why this matters more than it sounds like it should"

Most AI dev platforms implicitly optimize for one language ecosystem (usually JS/TS) because that's what's easiest to demo. That's fine until your actual stack doesn't match the demo. "Any language that fits in a container" means the platform adapts to your stack instead of the other way around — no rewrite, no adapter layer, no "well actually it only really works well in X."

"What this looks like in practice"

Say you've got a Go service doing something latency-sensitive sitting next to a Python service doing data processing. Instead of picking one language to "agentify" and leaving the other alone, both can live in the same Ardor workspace:

services/
  api-gateway/       # Go
  data-pipeline/     # Python
  Dockerfile         # per-service, whatever base image fits

Cerebrum builds and deploys each on its own terms, in its own container, without asking you to normalize your stack first.

"Honesty check"

This capability is currently listed as (Experimental) in Ardor's own docs — not because it doesn't work, but because it's early and the team would rather ship it and iterate in the open than oversell it as finished. If you try it and hit a rough edge, that tracks with where it actually is right now.