Skip to content
Developer docs

SDKs and CLI

What exists today for calling REM CloudHub from code: the OpenAPI spec for generated clients, the MCP server, the agent kit, and OpenAI-compatible AI calls.

REM CloudHub does not publish its own REST client library or command-line tool yet. Everything is reachable with plain HTTP, and these are the supported ways to work from code today:

Need

Use

A typed client in your language

Generate one from https://remcloudhub.com/api/openapi.json (OpenAPI 3.1) with any generator, such as openapi-generator or openapi-typescript

Shell scripts and CI

curl with a personal or service-account token; see REST API

An AI assistant that operates REM CloudHub

The MCP server

Agents in code

The agent kit (ADK), with its adk command

Model calls from your projects

Any OpenAI-compatible client pointed at your AI key's endpoint; see AI gateway, models and keys

A small deploy script

#!/usr/bin/env bash
# Redeploy a project from CI and wait for the result.
set -euo pipefail
API=https://remcloudhub.com/api
AUTH="Authorization: Bearer $RCH_TOKEN"     # a qchsa_ token bound to the apps area
APP_ID=118

curl -fsS -X POST -H "$AUTH" "$API/apps/$APP_ID/update-now"   # 202: "Updating <name> from <branch>"
sleep 5
for i in $(seq 1 60); do
  status=$(curl -fsS -H "$AUTH" "$API/apps/$APP_ID/deploys" | jq -r '.current.status')
  echo "status: $status"
  case "$status" in
    running) exit 0 ;;
    failed)  curl -fsS -H "$AUTH" "$API/apps/$APP_ID/deploys" | jq -r '.current.error'; exit 1 ;;
  esac
  sleep 10
done
echo "timed out"; exit 1

When a client library or CLI is released, it will be listed here.

Download: PDF · Word · Markdown