Your AI creates and updates projects, deploys source, and checks the result through MCP.
Your connected AI uses these MCP tools to carry out your request. Application HTTP calls are documented separately in API reference.
List existing projects first and use their IDs. Create a new project only when the user asks; project names are not unique.
list_projectsget_projectcreate_projectupdate_projectget_project_statusget_project_endpointsget_project_metricsUpload a source ZIP, wait for the build to finish, then check isLive and any warning before reporting success. Git push alone does not deploy.
create_buildget_buildlist_buildsupdate_buildcancel_buildrebuild_projectredeploy_buildCreate, change or pause a schedule, or run a project once. Read the run output before repeating a job that may have changed data.
list_schedulesadd_scheduleupdate_scheduledelete_schedulerun_projectRead runtime logs for the intended project and time range. Diagnose the recorded error before changing code or retrying an uncertain write.
get_logsEnvironment values take effect after a rebuild. Lists show names and setting status; read a secret value only when the user specifically asks. Domain setup still requires DNS changes at the domain provider.
list_environmentsget_environmentset_environmentdelete_environmentlist_firewallscreate_firewallupdate_firewalldelete_firewalllist_domainscreate_domaindelete_domainlist_project_managersset_project_managers