What does the license cover?
The Runner, signed updates, and support — entitlements checked at license verification. That check, plus downloads and docs, is all Cohenta’s cloud does; it is never in the execution path.
Who pays for model usage?
You do, through your own accounts. The Runner brings your own keys, so OpenAI, Anthropic, and your creative providers bill you directly at their rates. Cohenta never receives or stores your provider keys.
What do the free Skills actually do?
Campaign strategy, brand-voice assessment, repurposing plans, copy review — the thinking work, inside ChatGPT and Claude. They reason; the Runner executes.
How does approval work?
A run pauses at the gate and waits — durably, across restarts — until Approve or Request changes in Slack or Teams resumes that exact run. Publication needs its own explicit confirm, and ad creative always lands paused.
How is Cohenta Server licensed?
By Compute Unit — 1 CU = 4 vCPU — for production deployments on your own hardware: pinned to a node, or floating across a pool via a relay you run. Activation works fully air-gapped, the RS256-signed license verifies locally, and nothing phones home. It isn't self-serve checkout; the deployment is scoped on a call.
Why is a price missing right now?
The billing catalog didn’t answer when this page rendered, so rather than show a stale number the paid cards wait. The free Skills and the walkthrough don’t depend on it.