158 29
Text’s Model Context Protocol (MCP) server lets ChatGPT, Claude, and other compatible AI assistants access the tickets you’re authorized to see in Text App.
Once connected, you can find tickets, review tags, summarize customer cases, and analyze recurring issues using natural-language prompts. You can work with support data directly from your AI assistant without manually copying ticket content between applications.
What is MCP?
Model Context Protocol is an open standard that lets AI assistants connect securely to external tools and data sources.
Text hosts a remote MCP server that connects compatible AI assistants to your Text data. When you submit a request, the assistant selects an appropriate MCP tool, retrieves the relevant information, and uses it to prepare its response.
For example, instead of searching through your ticket list manually, you can ask:
Find tickets created this week that mention billing issues and summarize the most common problems.
The assistant can retrieve matching tickets and analyze them while respecting your existing Text permissions.
What can you do with Text’s MCP server?
You can use the MCP connection to:
-
Find a specific ticket by its ID.
-
List tickets using available filters.
-
Review tags associated with tickets.
-
Summarize long customer cases.
-
Find recurring questions or support problems.
-
Prepare briefings based on recent tickets.
-
Search and retrieve ticket information from ChatGPT or Claude.
The currently published MCP tools focus on finding, retrieving, and analyzing data. If you want to build a custom ticketing integration or automate actions, see Other ways to work with your tickets.
Before you connect
To connect an AI assistant, you need:
-
An active Text account.
-
Access to the relevant tickets and teams.
-
An AI assistant that supports the Text plugin or remote MCP connections.
-
A plan or workspace configuration that allows plugins or custom connections.
ChatGPT users can install the Text plugin without entering the MCP endpoint manually.
For Claude and other compatible MCP clients, use:
https://mcp.text.com/
OAuth is the recommended authentication method. During setup, you’ll be redirected to sign in to your Text account and authorize the connection.
Using HelpDesk without Text App? The MCP setup below applies to Text. To build an integration for your environment, use the HelpDesk API.
Connect ChatGPT to your tickets
The preferred way to connect ChatGPT is through the Text plugin:
-
Install the Text plugin in ChatGPT.
-
Log in with your Text account.
-
Mention @Text in the chat to ask questions about your data.
For example:
@Text List the five most recent tickets I can access.
After you mention @Text, ChatGPT can use the connection to retrieve relevant ticket information and prepare its answer.
Installing or using the plugin can depend on your ChatGPT plan, region, workspace settings, and role. In a managed workspace, an administrator may need to enable the plugin and its underlying app.
For more information, see Plugins in ChatGPT and Codex.
Connect Claude to your tickets
Remote MCP connectors are available in supported Claude plans. Organization owners may need to add or enable the connector before other members can use it.
To connect Text:
-
Open Claude and go to Settings.
-
Select Connectors.
-
If you use an organization account, open Organization connectors.
-
Select Add custom connector.
-
Enter Text as the connector name.
-
Enter the MCP server URL:
https://mcp.text.com/ -
Add the connector and select Connect.
-
Sign in to your Text account and approve the connection.
-
Open a conversation and use Search and tools to enable the Text connector.
Once enabled, Claude can select the appropriate Text tool when your request requires ticket information.
For current plan availability and connector controls, see Anthropic’s custom connector guide.
Test the connection
If you’re using ChatGPT, start with:
@Text List the five most recent tickets I can access.
If you’re using Claude or another connected assistant, try:
List the five most recent tickets I can access.
You can also test access to a specific ticket:
Find ticket 123456 and summarize the customer’s problem.
If the connection works, your assistant should show that it is using Text before returning the result. Depending on the assistant and its permissions, you may be asked to approve the request.
Available ticket tools
The tools available to you can depend on the AI assistant and connection method.
Tool | What it does |
|---|---|
| Finds an existing ticket by its ID. |
| Lists tickets with pagination and optional filtering. |
| Lists tags associated with a ticket. |
| Searches for relevant data in Text. |
| Retrieves specific resources for use in an AI-generated response. |
The available toolset may expand over time. Check the Text MCP documentation for the current list.
Prompts you can try
When using ChatGPT, mention @Text at the beginning of your prompt. In Claude, make sure the Text connector is enabled.
Find and summarize a ticket
Find ticket 123456. Summarize the customer’s problem, the actions already taken, and any unresolved questions.
Review urgent tickets
List the high-priority tickets assigned to my team and give me a short summary of each case.
Investigate recurring issues
Find tickets created this week that mention billing problems. Group them by issue and summarize the most common causes.
Analyze tickets with a specific tag
Review recent tickets tagged “integration” and identify the problems customers mention most frequently.
Prepare a support briefing
Prepare a briefing based on the tickets I can access that still require attention. Include the ticket ID, priority, main issue, and recommended next step.
Review ticket tags
Show me the tags associated with ticket 123456 and explain what they indicate about the case.
The quality and completeness of the response depend on your permissions, the selected tool, the filters used, and the amount of data requested.
How permissions work
The MCP server uses the permissions of the Text user who authenticated the connection.
This means:
-
Your assistant can only retrieve data you’re authorized to access.
-
Team assignments and user roles continue to apply.
-
Connecting an assistant doesn’t grant access to every ticket in the organization.
-
Different users may receive different results from the same prompt.
-
The assistant follows the authenticated user’s current roles, team access, and authorized resources.
Only connect through the official Text plugin or MCP endpoint and review the requested permissions before authorizing access.
You can manage or remove the Text plugin or connector from your AI assistant’s settings.
Other ways to work with your tickets
MCP is useful when you want to retrieve and analyze ticket data interactively from an AI assistant. Text also provides the Ticketing API and Skills for integrations and automation.
Text Ticketing API
Use the Text Ticketing API when you need to build a custom application or integration with direct, programmatic access to ticketing functionality.
The API is a better choice when your integration needs specialized business logic, operates independently of an AI assistant, or performs actions automatically from your application.
Start with the Ticketing API overview to understand how ticketing fits into Text’s messaging APIs. For available operations, parameters, and response structures, see the complete Ticketing API reference.
Skills
Use Skills when you want to automate workflows directly in Text.
Skills can combine:
-
Predefined triggers and actions.
-
Ticket and conversation data.
-
Conditions and multiple workflow paths.
-
AI for interpreting, generating, or summarizing content.
-
Webhooks for connecting Text with external services and APIs.
Skills are a good choice when you want an automation to respond to a customer interaction or support event without building and hosting a complete custom integration.
For example, a Skill can collect information, use AI to interpret the request, perform predefined ticket actions, and send selected data to an external system through a webhook.
Troubleshooting
The Text plugin isn’t available in ChatGPT
Installing and using a plugin can depend on your plan, region, workspace settings, role, and supported ChatGPT surface.
If you use a managed workspace, ask an administrator to confirm that the Text plugin and its underlying app are enabled for your role.
ChatGPT doesn’t use Text in its response
Make sure the Text plugin is installed and connected to the correct Text account. Mention @Text explicitly at the beginning of your prompt.
Authentication fails
Confirm that you’re signing in to the correct Text account and that you approve the requested permissions during the authentication process.
The assistant can’t find a ticket
Make sure the authenticated Text user has access to the ticket and its assigned team. The connection can’t bypass Text roles, team assignments, or resource permissions.
Also check that the ticket ID and any filters used in the prompt are correct.
The response is incomplete or takes too long
Try narrowing the request. Add a date range, tag, priority, team, or another filter instead of asking the assistant to analyze a large ticket collection at once.
Frequently asked questions
Can ChatGPT access my tickets?
Yes. Install the Text plugin, log in with your Text account, and mention @Text in the chat when you want ChatGPT to use your ticket data.
Can Claude access my tickets?
Yes. Add Text as a custom connector in Claude, authenticate with your Text account, and enable the connector from Search and tools.
Can the AI assistant see every ticket?
No. The assistant can only access tickets available to the authenticated Text user. Existing roles, teams, and resource permissions continue to apply.
Can MCP create or update tickets?
The currently published external MCP ticket tools focus on finding and retrieving ticket data. For custom ticket operations and automated processes, consider the Text Ticketing API or Skills.
When should I use the Ticketing API instead of MCP?
Use MCP when you want to work with ticket information interactively through an AI assistant. Use the Ticketing API when you’re building a custom application or service that requires direct programmatic access to ticketing functionality.
When should I use Skills?
Use Skills when you want to automate a workflow in Text using predefined actions, AI, conditions, and webhooks without building a complete custom application.
Can I use another AI assistant?
Yes, provided that it supports compatible remote MCP servers and the authentication method used by Text. Setup steps can differ between assistants, so check the provider’s current MCP documentation.
