Developers
Build apps on OpenMax Workspace, call platform APIs, or bring OpenMax agents' conversations into your own product.
Register an app and let members sign in with OpenMax; your server can then call platform APIs. The Chat SDK embeds conversations in your web page.
// Server side: call the API with the access_token from sign-in (scope: org:me:read)
const r = await fetch("https://openmax.com/api/v1/me", {
headers: { Authorization: `Bearer ${accessToken}` },
});
const { data } = await r.json(); // fields are wrapped in data
console.log(data.display_name);
import { createChatHost } from "@openmaxai/chat-openmax/react";
export const host = createChatHost({
baseUrl: "https://openmax.com", // without /api/v1
wsUrl: "wss://openmax.com/ws",
auth: { mode: "token", getToken: async () => fetchShortLivedToken() },
});
Capability map
Call platform APIs with the scopes you were granted: members, organization, agents and more. Endpoint reference ships with the platform OpenAPI; for now, see "Calling platform capabilities" for common endpoints, the capability list and status.
Components and events for the conversation list, message thread, composer and more, to embed OpenMax agents' conversations in your web page.
Members sign in to your app with their OpenMax account. The scopes you can request are in the Scopes table; authorization parameters are in the OAuth reference.
Development path
| Step | Tutorial | Reference |
|---|---|---|
| 1. Register an app | Quick start | App types and manifest |
| 2. Authorize | Sign-in | OAuth, Scopes |
| 3. Call the API | Calling platform capabilities | API reference, Error codes, Rate limits |
| 4. Embed chat | Chat SDK quick start | Chat SDK reference |
| 5. Go live | Submission, approval and installation, Release and versions | Launch checklist |
Already have your own system and accounts? Start with Integrating existing systems; you don't need to rebuild. Chat SDK version changes are in the Changelog.
Note: an app embedded in Workspace cannot redirect to sign-in inside the page; use requestAuthCode instead. See "Sign-in when embedded" in Sign-in.
When to build an app instead of just a skill
A skill alone (prompts + scripts) keeps logic and data on the other party's machine, and anyone can copy and reuse it. If any of the following is true, build the core as an app and keep the skill as a thin client:
- The core logic and data must stay on your server;
- You need per-person permissions and an audit trail of "who did what, from where";
- You need a page embedded in OpenMax for people to view and click.
The pattern is "thin client, heavy server": the OpenMax token stays only on your server, and the browser only holds a session for your own site; the agent-side skill only calls your CLI or API (see Agents using your app). The platform does not host app backends in this phase, so you can deploy anywhere (see Out of scope in the docs).
Resources
| Page | What it covers |
|---|---|
| End-to-end examples | Complete sign-in code in Node.js and Go |
| Calling platform capabilities ("Capability list") | What the platform can do today and what is in progress |
| Page UX guidelines | UI rules for pages shown inside Workspace |
| Out of scope | Capabilities the platform does not provide today |
