The Text Model Context Protocol (MCP) is a standardized and secure way to connect AI Assistants with your Text App. It grants assistants access to tickets, archived chat transcripts, team resources, and metadata such as tags.
By integrating with the MCP Server, you can build intelligent, secure workflows in which your assistant can answer questions, retrieve data, and assist in support operations—all while respecting your team’s access controls.
Key Benefits
-
AI-Powered Productivity: Use AI Assistants to search and analyze tickets and chats.
-
Secure and Scoped Access: Permissions mirror your access in the Text platform - no overreach, no leaks.
-
Easy Configuration: Recommended OAuth-based setup for quick and secure connection; JSON configuration and API tokens are also supported for advanced users.
-
Curated Tools: Purpose-built endpoints allow assistants to retrieve data efficiently.
Available Integration Tools
Ticket Operations:
-
find-tickets
- Find tickets by ID, Agent, or tag -
list-tickets
- List tickets with filtering and pagination -
list-ticket-tags
- View available ticket tags
Chat Operations:
-
list-archived-chats
- Access chat history with filtering -
get-chat-transcript
- Retrieve full conversation transcripts -
list-chat-tags
- View available chat tags
Deep Research Integration (ChatGPT only):
-
search
- Search across tickets and chats -
fetch
- Retrieve relevant data for context-enriched responses
Security Approach
The MCP integration is designed with security in mind. Your data stays protected, and access is always aligned with your existing permissions and organizational structure.
-
Permission-Aware Access: Every request made through the AI assistant adheres to your existing user permissions and team assignments.
-
Strict Data Isolation: Users can only view data they are explicitly authorized to access.
Setup Guide
Prerequisites:
-
Active Text platform account with correct permissions.
-
AI assistant supporting MCP integration (e.g., Claude)
-
Basic JSON configuration experience
Configuration Steps:
-
Open your AI assistant's integrations panel.
-
Add new integration using the MCP Server endpoint:
https://mcp.text.com/
-
Choose your authentication method:
A. OAuth (Recommended): Browser-based, secure
B. Bearer Token: Use a direct API token. -
Test the connection to confirm tool availability.
-
Start issuing queries using available tools and natural language prompts.
Assistant-Specific Setup Resources:
-
Claude Desktop - MCP setup documentation
-
Claude Code - MCP setup documentation
-
OpenAI (ChatGPT) - MCP setup documentation
-
Claude Web - Go to settings > Integrations > + Add integration, then use https://mcp.text.com/mcp
-
Cursor - MCP setup documentation -> Coming soon!
-
Windsurf - MCP setup documentation -> Coming soon!
-
VS Code - MCP setup documentation -> Coming soon!
Usage Examples
Try the following prompts with your assistant:
"Show me all high-priority tickets assigned to my team."
"Get the transcript from yesterday's chat with customer@company.com."
"Find tickets created this week that mention billing issues."
"Analyze chat transcripts for common customer pain points."
These queries will automatically invoke the correct tools (for example, list-tickets
, get-chat-transcript
, or search
) under the hood.
Troubleshooting
If you run into issues while setting up or using the MCP integration, here are some common areas to check.
Start by verifying your connection. Ensure that the MCP Server endpoint URL (https://mcp.text.com/
) is correctly entered in your AI assistant’s configuration settings, and that your network allows outbound HTTPS connections to this address.
If you're having authentication problems, double-check your credentials.
When you encounter errors accessing certain tools (like get-chat-transcript or list-tickets), the issue often lies with user scope or team assignments. The MCP Server enforces Text’s permission system, so assistants can only retrieve data that the authenticated user is authorized to see. Make sure your account has the necessary access rights in the Text platform.
Finally, if you notice performance delays or partial results, be aware that the API has rate limits and best practices for bulk operations. Try narrowing your query filters or batching large requests for more efficient processing.