Using ChatGPT, Claude, and Other AI Tools

In addition to the AI Operator built into the Telerivet web app, you can use Telerivet from other AI assistants and agents, such as ChatGPT, Claude, Claude Code, and Codex.

These AI tools connect to Telerivet using MCP (Model Context Protocol), an open standard that lets AI assistants securely use other applications. Once connected, you can ask your AI assistant to work with your Telerivet account, for example:

• "Send a reminder to everyone in the Volunteers group about tomorrow's meeting"
• "How many messages failed to send last week, and why?"
• "Find contacts who haven't replied in 90 days and add them to a new group"
• "Create a poll that asks customers to rate our service from 0 to 10"
• "Show me my message statistics and costs for each route this month"

This lets you combine Telerivet with everything else your AI assistant can do, such as drafting messages based on a document you're working on, or analyzing your messaging data alongside data from other sources.

For example, here is ChatGPT creating a poll in Telerivet:

Connecting an AI Assistant

Telerivet provides setup instructions for these AI assistants:

• ChatGPT
• Claude (claude.ai and the Claude desktop app)
• Claude Code
• Codex
• Antigravity CLI
• Hermes Agent

Other AI tools that support MCP can connect using Telerivet's MCP server URL, https://api.telerivet.com/mcp. See the MCP documentation for technical details.

Authorizing Access

When you connect an AI assistant, it opens a Telerivet page where you log in and authorize access. On this page, you choose:

• Permissions – which types of data and actions the AI assistant can access, such as viewing messages, sending messages, or editing automated services. Uncheck any permission you don't want it to have. The AI assistant can never do more than your own Telerivet account has permission to do.
• Organization – which of your Telerivet organizations the AI assistant can access. To use it with another organization, connect it again.

Each AI assistant must be approved by an organization administrator before it can be used in an organization. If you are an organization administrator, the authorization page lets you approve the AI assistant for your organization, and choose which users can use it. Other users will see a message asking them to contact an organization administrator.

Approving Actions

By default, the AI assistant needs your approval before it can send messages, trigger automated services, or delete data. Depending on the AI assistant, you can approve an action in one of these ways:

• Approval prompt – The AI assistant asks you to approve the action directly, with a description of what the action will do. (Approval prompts are currently supported in Claude Code.)
  
  

• Message draft – The AI assistant shows an editable draft of the message and its recipients. You can edit the From route and message content before clicking Send to approve sending that exact message. Alternatively, you can click Edit in Telerivet to open the Telerivet web app in a separate tab, which allows editing additional message options such as attachments, short links, labels, and schedule. (Editable drafts are currently available in ChatGPT and Claude when sending simple text messages.)

  

• Approval link – The AI assistant gives you a link to an Approve Action page in Telerivet, which shows the details of the action. After clicking Approve, return to the AI assistant and ask it to try again. (Approval links are used whenever approval prompts and message drafts are not supported.)
  
  
  When the user clicks the approval link ("Review and approve it here"), the Approve Action page is opened in a separate tab:
  
  

Each approval applies only to one exact action, and expires after 1 hour. If Message Review is enabled for your project, messages sent by AI assistants also follow the same review rules as messages sent from the Telerivet web app.

Controlling AI Assistants in Your Organization

Organization administrators can control what each AI assistant is allowed to do in their organization on the Connected Applications page (also linked from your organization's Developer API page):

Click an application to view or change its settings:

• Required permission – which users in the organization can connect the AI assistant: all users, organization administrators only, or organization and project administrators only.
• Allowed actions – for each type of action (reading data, creating and updating data, sending messages and triggering services, deleting data, and making external web requests), whether the AI assistant can perform it on its own (Allowed), must ask a user first (Require approval), or can't perform it at all (Not allowed).
• Allowed permissions – the types of data and actions the AI assistant can access in the organization, even if a user granted more permissions when connecting it.
• Revoke access – removes the AI assistant from the organization and disconnects it for all users.

Disconnecting an AI Assistant

You can disconnect an AI assistant from your own Telerivet user account at any time in the Authorized Applications section of your Account Security settings. Organization administrators can also revoke access for any user on the OAuth Access Tokens page.