Developer overview
Everything a program can do in REM CloudHub: the REST API, the MCP server, tokens, webhooks and the agent kit, and where each one lives.
Everything you can do in the REM CloudHub console, a program can do too, with the same permissions as the person or service account behind its token.
Interface | Address | Use it for |
|---|---|---|
REST API |
| Scripts, CI pipelines, your own tools. About 1,500 operations |
OpenAPI spec |
| Generate a client in any language; explore it at |
MCP server |
| AI clients (Claude, IDE assistants, your own agents) acting in REM CloudHub with your rights |
AI gateway | the endpoint on your key's Use it tab | OpenAI-compatible model calls with a REM CloudHub AI key |
Keyguard |
| Calling outside APIs without holding their keys |
Webhooks | inbound and outbound | Deploy on push, start agents, receive alerts |
Agent kit (ADK) | a Python package | Building agents in code that REM CloudHub runs |
Start here
Authentication and tokens: personal tokens (
qchp_), service-account tokens (qchsa_), binding, and the organisation header.REST API: the main endpoints with examples.
MCP server: connect Claude or another MCP client.
Service accounts: credentials for unattended programs.
Webhooks: push-to-deploy, agent triggers and alert deliveries.
Agent kit (ADK): write an agent in one file.
Errors: the error body and request references.
SDKs and CLI: what exists today.
Ground rules
Every call is made as someone: the person who owns the token, or the service account. A token never has more rights than its owner, and it can be narrowed further.
Every change is recorded in the audit trail with the token that made it.
Lists answer for one organisation at a time: the one named in the
X-RCH-Orgheader, or your personal organisation.Secret values are write-only through every interface: you can set and rotate them, never read them back.