AI Power for Request Tracker

✦ Now includes RT MCP Server

AI Power for Request Tracker,Your Way

Suggested replies, ticket summaries, sentiment detection, smart routing, and more, right inside RT. Plus the new RT MCP Server lets you talk to your tickets directly from Claude or any AI assistant.

🔌 Bring your own AI engine
🔒 Data stays on your infrastructure
☁️ Cloud & on-prem RT 6
⚙️ Fully open source
RT MCP Server

Your AI assistant can now actually see your tickets

Claude and other AI tools are great at reasoning, but they’ve never had a way to look at your actual tickets. The open-source RT MCP Server changes that. Point it at your RT instance and start asking questions in natural language.

  • Ask about tickets in natural language, no RT query syntax required
  • Get the full picture: history, comments, status, custom fields, all of it
  • Update tickets, reassign work, and change due dates without leaving your chat
  • Leave comments directly from your AI session
  • Everything goes through your own RT API. Nothing touches outside servers
Claude + RT MCP Server
You › What high-priority tickets have been open longest?

Querying RT… 3 tickets unresolved > 72 hours:

#4821 Payment gateway timeouts – 4 days
#4799 SSO intermittent failures – 6 days
#4776 Report export crash on large data – 8 days

You › Assign #4776 to alex, due Friday, add a comment we’re investigating

✓ Assigned to [email protected]
✓ Due date set to Friday
✓ Comment added to ticket #4776

See It In Action

Watch RT MCP in action

Watch how AI transforms ticket handling in Request Tracker, from summaries and smart replies to a full conversational interface for your queue.

Terminal Workflow

Using the RT MCP Server from a Claude Code terminal session.

Desktop App

Working with RT through Claude’s desktop application. No terminal required.

Capabilities

The features that actually help

Quick summaries, smarter replies, sentiment flags. These sit right inside RT and fit into how your team already works.

📋

Ticket Summaries

Long thread? One click gets you a plain-language summary so anyone can jump in without reading 40 replies.

✍️

Smart Autocomplete

As you type a reply, AI suggests how to finish it. Accept, tweak, or ignore. You’re always in control.

Inline AI Actions

Select any message and rewrite it, translate it, or soften the tone. One click, and it’s handy for global teams.

😟

Sentiment Analysis

Flags tickets where the customer sounds frustrated, before things escalate, so you get to the right ones first.

🔌

RT MCP Server

Connect your AI tool directly to RT. Ask questions, take action, leave comments, all from your AI chat.

New · Open Source

🔒

Private by Design

Your ticket data only ever goes through your own RT instance, whether you’re using the extension or MCP Server.

Security

🤖

Bring Your Own AI Engine

Hook up OpenAI, Gemini, Claude, or a self-hosted model. No proprietary AI subscription, just your own API key.

🎭

Customizable AI Behavior

Adjust prompts, tone, and which features are on, per queue or globally.

🗂️

Smart Routing

New ticket come in? AI suggests the right queue, owner, and priority based on content.

🔍

Similar Ticket Lookup

RT surfaces related tickets already resolved while you compose, so you’re not writing the same answer twice.

Flexibility

Use the AI engine you prefer

Both tools are vendor-agnostic. Bring your own API key or run something locally. No lock-in, no separate RT AI subscription required.

OpenAI / GPT-4o
Google Gemini
Anthropic Claude
Custom / Self-hosted

Compatible with RT 6 · Cloud or on-premises deployments

Why It’s Different

Built to fit how you already work

RT’s AI doesn’t try to replace your judgment. It plugs into how your team already uses RT: ticket creation, triage, responses, and comments.

With the RT MCP Server, your AI assistant works with RT in real time, from wherever you already have your AI chat open.

  • Bring Your Own AI: OpenAI, Gemini, Claude, or custom models
  • Fully open source and extendable
  • Runs on your RT instance, cloud or on-prem
  • Built to plug into RT 6 with minimal setup
  • RT MCP Server: conversational access from any MCP-compatible tool
  • Data never leaves your own infrastructure
Getting Started

Set up RT MCP in minutes

The RT MCP Server uses RT’s existing REST API, so there’s nothing extra to install in RT itself. All you need is a URL and an auth token.

1

Install the MCP server

Clone the open-source repo and run npm install. No RT plugin required.

2

Configure your connection

Add your RT instance URL and API token to a .mcp.json config file. Your credentials stay on your own machine.

3

Connect your AI assistant

Point Claude or another MCP-compatible tool at the server. Works in terminal and desktop AI apps including Claude Desktop.

4

Start talking to RT

Ask questions, update tickets, reassign work, leave comments, all in natural language, without leaving your AI chat.

Want to see it for yourself?

We’re still early days with AI in RT and adding to it regularly. If you have thoughts on what would actually help your team, we’d genuinely love to hear them.