Skip to main content
SmartAlex is built for AI-native builders, and that includes ChatGPT. Add the hosted SmartAlex connector and ChatGPT can operate a full voice AI platform in plain conversation: spin up a voice agent from a website URL, launch an outbound campaign, read call transcripts, and move a deal to closed_won, without you leaving the chat. The connection runs over the Model Context Protocol (MCP). One hosted address, a sign-in to approve, and ChatGPT gets a set of tools it can act with, not just documentation to read. That is the difference between a chatbot that can describe your account and one that can run it.
Honest framing: SmartAlex is agent-native by design and our MCP server is live and listed in the ChatGPT and Claude connector directories. It works the same way from Cursor, VS Code, Windsurf, Claude Code, and Gemini CLI. This is not an exclusive or official partnership with OpenAI or any other client vendor. ChatGPT is simply one of the smoothest ways in.

What ChatGPT can do once connected

Connected, ChatGPT drives the platform through the SmartAlex MCP server: 80 tools (all prefixed smartalex_). 31 only read, 30 create or change data, and 19 ask you first because they spend money, contact real people or delete something for good. The main areas: Tools your plan, add-ons or role don’t allow are hidden. Campaign tools and buying numbers need the Outbound Campaigns add-on. HTTP tools and creating or deleting webhooks need Power Tools. Creating an agent needs a free agent slot. The headline tool, smartalex_create_agent, takes a website URL, reads the business, generates an optimized prompt, and deploys a live voice agent in about 30 to 60 seconds. Underneath, SmartAlex runs a curated, high-accuracy speech, language, and voice stack, kept abstracted, so there is nothing to tune.

Before you connect

You need two things: a ChatGPT plan that supports custom connectors (with developer mode enabled where required), and a SmartAlex account. ChatGPT acts as you, with your role’s permissions. If you cannot see the connector option in ChatGPT, check the plan and developer mode. If you sign in but some tools are missing, check your plan, add-ons and role. See Security and permissions.

Connect ChatGPT to SmartAlex

The path is the hosted server address and a sign-in: no install, no key to paste, no config file. The address is the same one every MCP client uses:
1

Open connector settings in ChatGPT

Go to Settings, then Connectors (custom connectors). On plans that require it, turn on developer mode first. Choose Add or Create a custom connector.
2

Add SmartAlex

Paste the address above and name it SmartAlex. Because we are listed in the ChatGPT connector directory, you may also be able to find SmartAlex by name instead of pasting the URL. Transport is HTTP Streamable, authentication is OAuth 2.0 (PKCE, S256).
3

Sign in and allow

ChatGPT opens a sign-in page. Sign in to SmartAlex and press Allow. ChatGPT works in the workspace you last opened in SmartAlex, with your role’s permissions. Nothing to copy back.
4

Ask it to do something

Start a chat and type a plain request such as “show me my SmartAlex agents”. ChatGPT selects the right tool and runs it against your workspace.

Example prompts

Real requests you can paste into ChatGPT once the connector is live:
Actions that spend money, contact real people or delete something for good ask you first. For a request like “delete the old test agent”, ChatGPT gets a preview of exactly what will happen and must show it to you and wait for your yes. It can then run the action only with a one-time code from that preview, which works once and expires after 10 minutes.

What about GPT Actions?

If you are building a custom GPT and reaching for GPT Actions to make a voice agent, here is the honest state of play. GPT Actions call an OpenAPI-described REST endpoint. Our public REST API and TypeScript SDK are in private beta (“coming soon”), so the Actions path is not the way in yet. The live, supported route today is the MCP connector above, which already exposes the full 80-tool operating surface to ChatGPT. When the REST API ships, a GPT Actions integration becomes possible too, and we will document it here.

Embed the result on your site

When an agent is ready, ask ChatGPT for the widget code (smartalex_get_widget_install_code) and drop it in. For a static site, paste this single tag just before the closing body tag:
For a React or Next.js app, inject it from your root layout:
The snippet is byte-identical for every widget style: the widget reads its published config (variant, colors, agent) from the server at load time, so you can restyle it in Widget Studio later without touching the site again. YOUR_TENANT_ID is shown in the widget code panel inside SmartAlex, or returned by smartalex_get_widget_install_code.

Live today, coming soon

Live now

The hosted MCP server (connect by signing in) and Custom HTTP Tools are in production. These are the programmatic surfaces ChatGPT builds against today.

Coming soon

A public REST API and TypeScript SDK (which would unlock GPT Actions) are in private beta. Outbound webhook event delivery is also in private beta.

MCP getting started

Connect ChatGPT, Claude Code, Cursor, VS Code, or Gemini CLI in under two minutes.

What an AI agent can do

The 80 tools, what asks you first, and the per-workspace isolation model.

Build with Claude Code

The same MCP server, driven from the terminal.

Pricing and add-ons

Plans, included minutes, and add-ons.