With a Jira MCP server, AI models or local MCP clients (such as Claude Desktop) can perform Jira operations programmatically while adhering to user permissions and project security rules. This setup allows natural language commands to be translated into actionable Jira API requests and responses. Through this setup, MCP servers act as intelligent intermediaries that:<\/p>\n\n\n\n
- \n
- Retrieve and modify Jira data such as issues, sprints, and projects,<\/li>\n\n\n\n
- Run JQL queries for task reports, and<\/li>\n\n\n\n
- Execute multi-step project management workflows without leaving the AI environment.<\/li>\n<\/ul>\n\n\n\n
Learn more about MCP and how it enables enterprise-ready AI integrations<\/em><\/a>.<\/em><\/p>\n\n\n\n
![The Jira MCP integration works as a middleware layer between AI systems and Jira\u2019s REST API](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-25-1.png\")
<\/a><\/figure>\n\n\n\nHow Does Jira MCP Work?<\/h2>\n\n\n\n
The Jira MCP integration works as a middleware layer between AI systems and Jira\u2019s REST API. It translates natural language requests from AI models into API calls and then formats Jira\u2019s responses back into structured outputs that the model can interpret.<\/p>\n\n\n\n
Here\u2019s what the workflow looks like.<\/p>\n\n\n\n
- \n
- Connection and Authentication<\/li>\n<\/ol>\n\n\n\n
The MCP server connects to Jira using secure credentials, typically an API token or OAuth 2.0 authentication. This ensures that all operations performed by AI agents comply with Jira\u2019s access controls and project permissions<\/a>.<\/p>\n\n\n\n
- \n
- Request Initiation<\/li>\n<\/ol>\n\n\n\n
An AI agent<\/a> (for example, Claude Desktop or another MCP-compatible client) sends a natural language instruction such as:<\/p>\n\n\n\n
\u201cCreate a bug in the frontend project with the title \u2018Login button not responding.\u2019\u201d<\/em><\/p>\n\n\n\n
- \n
- Interpretation and Validation<\/li>\n<\/ol>\n\n\n\n
The MCP server receives the request, validates the agent\u2019s identity or token, and checks any required permissions or role-based access controls (RBAC) before executing the action.<\/p>\n\n\n\n
- \n
- Translation to API Call<\/li>\n<\/ol>\n\n\n\n
The server converts the natural language instruction into a structured Jira API request. For example, a POST \/rest\/api\/3\/issue<\/strong> call with all required parameters (project key, summary, description, and issue type).<\/p>\n\n\n\n
- \n
- Execution<\/li>\n<\/ol>\n\n\n\n
The MCP server takes that raw API response and reformats it into a concise, machine-readable or human-friendly output \u2014 typically JSON or plain text \u2014 and returns it to the AI agent.<\/p>\n\n\n\n
How Jira Integrates with MCP<\/h2>\n\n\n\n
Jira MCP integration works through authenticated API communication. The MCP server connects to Jira using API tokens or OAuth credentials. This allows AI agents to fetch or update project data securely. Integrating Jira with MCP can be done in multiple ways, depending on your infrastructure and automation needs.<\/p>\n\n\n\n
- \n
- Using Workato\u2019s pre-spun Jira MCP instance<\/li>\n<\/ol>\n\n\n\n
Workato provides a secure<\/a>, managed MCP instance pre-spun and pre-configured for Jira ready to use. You can connect to it instantly without deploying your own server. See the Workato Agentic MCP<\/a> product page. It\u2019s suited for enterprises needing governance<\/a> and speed.<\/p>\n\n\n\n
- \n
- Building your own self-hosted MCP server (Node.js example in this guide)<\/li>\n<\/ol>\n\n\n\n
You can self-host an MCP server using Node.js, Python, Go, or Java. This gives you full control and more customization but requires additional setup and maintenance. @modelcontextprotocol\/sdk + <\/strong>jira.js<\/strong><\/a> (the official client for the Jira REST API v3) for Jira API calls is the recommended stack for this integration.<\/p>\n\n\n\n
- \n
- Dockerized MCP<\/li>\n<\/ol>\n\n\n\n
You can run the MCP in an isolated container for production environments.<\/p>\n\n\n\n
- \n
- Hybrid setup<\/li>\n<\/ol>\n\n\n\n
In a hybrid setup, you can combine Workato\u2019s managed Enterprise MCP<\/a> environment with self-hosted connectors for specific custom logic.<\/p>\n\n\n\n
In this guide, we\u2019ll focus on building a Node.js MCP server and connecting it to Jira step by step.<\/p>\n\n\n\n
Getting Started with Jira MCP: Setting Up Your Jira MCP Server<\/h2>\n\n\n\n
This section provides an end-to-end walkthrough for setting up a Jira MCP server using Node.js and configuring it for the Claude Desktop integration as your connected AI agent. Let\u2019s get started.<\/p>\n\n\n\n
Step 1: Prerequisites<\/h3>\n\n\n\n
Before you begin, ensure you have the following:<\/p>\n\n\n\n
- \n
- Node.js v20.x or higher<\/li>\n\n\n\n
- npm or yarn<\/li>\n\n\n\n
- A Jira Cloud account (or Jira Data Center with API access)<\/li>\n\n\n\n
- Jira API token<\/li>\n\n\n\n
- Claude Desktop (for client testing)<\/li>\n<\/ul>\n\n\n\n
Check Node.js version:<\/p>\n\n\n\n
node –version<\/p>\n\n\n\n
If missing, download it from <\/a>nodejs.org<\/a>.<\/p>\n\n\n\n
Step 2: Authentication Setup<\/h3>\n\n\n\n
For Jira Cloud:<\/p>\n\n\n\n
- \n
- Visit: https:\/\/id.atlassian.com\/manage-profile\/security\/api-tokens<\/a><\/li>\n\n\n\n
- Click \u201cCreate API token\u201d<\/strong><\/li>\n\n\n\n
- Name it \u201cJira-MCP-Integration\u201d<\/li>\n\n\n\n
- Copy and securely store your token<\/li>\n<\/ol>\n\n\n\n
You\u2019ll use it to authenticate your MCP server with Jira.<\/p>\n\n\n\n
For Jira Server or Data Center:<\/p>\n\n\n\n
Navigate to your Jira profile \u2192 Personal Access Tokens \u2192 Create Token<\/strong> \u2192 set a name and expiry, then copy it.<\/p>\n\n\n\n
Step 3: Installation<\/h3>\n\n\n\n
Create a new project using these commands:<\/p>\n\n\n\n
mkdir jira-mcp-server<\/p>\n\n\n\n
cd jira-mcp-server<\/p>\n\n\n\n
npm init -y<\/p>\n\n\n\n
Install dependencies:<\/p>\n\n\n\n
npm install @modelcontextprotocol\/sdk jira.js dotenv p-retry<\/p>\n\n\n\n
This command does the following:<\/p>\n\n\n\n
- \n
- @modelcontextprotocol\/sdk<\/strong>: defines MCP server and tool APIs.<\/li>\n\n\n\n
- jira.js<\/strong>: official Jira client supporting REST API v3.<\/li>\n\n\n\n
- dotenv<\/strong>: loads .env<\/strong> into process.env<\/strong>.<\/li>\n\n\n\n
- p-retry: <\/strong>helper for retrying transient API errors (e.g., rate limits and network blips).<\/li>\n<\/ol>\n\n\n\n
Output:<\/strong><\/p>\n\n\n\n
added 47 packages in 2.6s<\/p>\n\n\n\n
Check your package.json<\/strong>:<\/p>\n\n\n\n
{
“name”: “jira-mcp-server”,
“version”: “1.0.0”,
“type”: “module”,
“scripts”: {
“start”: “node index.js”,
“test”: “node test\/test.js”
},
“dependencies”: {
“@modelcontextprotocol\/sdk”: “^1.21.0”,
“jira.js”: “^5.2.2”,
“dotenv”: “^17.2.3”,
“p-retry”: “^5.0.0”
}
}<\/p>\n\n\n\nAdd “type”: “module” <\/strong>to package.json<\/strong> to use ES modules (imports).<\/p>\n\n\n\n
Step 4: Configure Environment Variables<\/h3>\n\n\n\n
Create a .env<\/strong> file to store Jira credentials:<\/p>\n\n\n\n
touch .env<\/p>\n\n\n\n
Add:<\/p>\n\n\n\n
JIRA_HOST=”your-domain.atlassian.net”<\/p>\n\n\n\n
JIRA_EMAIL=”you@example.com”<\/p>\n\n\n\n
JIRA_API_TOKEN=”your_api_token_here”<\/p>\n\n\n\n
Your MCP server will read these variables for secure authentication. Don\u2019t forget to add your .env<\/strong> to .gitignore<\/strong>.<\/p>\n\n\n\n
Step 5: Initialize the MCP Server<\/h3>\n\n\n\n
Create a file named index.js<\/strong> and add the following JavaScript code to initialize your MCP Server:<\/p>\n\n\n\n
import ‘dotenv\/config’;
import { McpServer } from ‘@modelcontextprotocol\/sdk\/server\/mcp.js’;
import { StdioServerTransport } from ‘@modelcontextprotocol\/sdk\/server\/stdio.js’;
import { Version3Client } from ‘jira.js’;<\/p>\n\n\n\n\/\/ Initialize Jira client
const jira = new Version3Client({
host: process.env.JIRA_HOST,
authentication: {
basic: {
email: process.env.JIRA_EMAIL,
apiToken: process.env.JIRA_API_TOKEN,
},
},
});<\/p>\n\n\n\n\/\/ Create the MCP server
const server = new McpServer({
name: ‘jira-mcp’,
version: ‘1.0.0’,
});<\/p>\n\n\n\n\/\/ Tool: Get issues by JQL
server.tool(
‘getIssuesByJQL’,
‘Fetch Jira issues using a JQL query’,
{
jql: { type: ‘string’, description: ‘JQL string (e.g., project = TEST)’ },
maxResults: { type: ‘number’, description: ‘Limit results’, default: 50 },
},
async ({ jql, maxResults }) => {
console.error(`Running JQL: ${jql}`);
const response = await jira.issueSearch.searchForIssuesUsingJql({ jql, maxResults });
console.error(`Found ${response.issues?.length || 0} issues`);
return {
content: [{ type: ‘text’, text: JSON.stringify(response.issues, null, 2) }],
};
}
);<\/p>\n\n\n\n\/\/ Tool: Create new Jira issue
server.tool(
‘createIssue’,
‘Create a new Jira issue’,
{
projectKey: { type: ‘string’, description: ‘Project key (e.g., TEST)’ },
summary: { type: ‘string’, description: ‘Issue title’ },
description: { type: ‘string’, description: ‘Issue details’ },
issueType: { type: ‘string’, description: ‘Type (Task, Bug, etc.)’ },
},
async ({ projectKey, summary, description, issueType = ‘Task’ }) => {
console.error(`Creating issue in ${projectKey}`);
const issue = await jira.issues.createIssue({
fields: {
project: { key: projectKey },
summary,
description,
issuetype: { name: issueType },
},
});
console.error(`Created issue: ${issue.key}`);
return {
content: [{ type: ‘text’, text: JSON.stringify(issue, null, 2) }],
};
}
);<\/p>\n\n\n\n\/\/ Start the MCP server
async function main() {
const transport = new StdioServerTransport();
await server.connect(transport);
console.error(‘Jira MCP Server is running’);
}
main().catch((err) => console.error(‘Error:’, err));<\/p>\n\n\n\nThis JavaScript code snippet does the following:<\/p>\n\n\n\n
- \n
- The getIssuesByJQL<\/strong> tool fetches issues via a JQL query.<\/li>\n\n\n\n
- The createIssue<\/strong> tool allows you to add new issues from Claude or any MCP client.<\/li>\n\n\n\n
- McpServer.tool(…)<\/strong> defines callable tools with typed inputs. Clients map intents to tool calls.<\/li>\n\n\n\n
- StdioServerTransport<\/strong> is convenient for desktop clients (e.g., Claude Desktop) that spawn the MCP server process and communicate over stdio. For remote deployments, use socket\/HTTP transports supported by the MCP SDK.<\/li>\n<\/ol>\n\n\n\n
Step 6: Run Your Jira MCP Server<\/h3>\n\n\n\n
Start it:<\/p>\n\n\n\n
node index.js<\/p>\n\n\n\n
#or<\/p>\n\n\n\n
npm test<\/p>\n\n\n\n
Now your Jira MCP server is running:<\/p>\n\n\n\n
![Jira MCP server is running](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-31-1-1024x321.png\")
<\/a><\/figure>\n\n\n\nStep 7: Configure Claude Desktop as your Connected AI Agent<\/h3>\n\n\n\n
Once your MCP server is running, you can configure your AI agent to use the Jira MCP integration.<\/p>\n\n\n\n
First, locate the Claude configuration file as follows:<\/p>\n\n\n\n
- \n
- macOS path: ~\/Library\/Application Support\/Claude\/claude_desktop_config.json<\/strong><\/li>\n\n\n\n
- Windows: %APPDATA%Claudeclaude_desktop_config.json<\/strong><\/li>\n<\/ol>\n\n\n\n
![Local MCP server settings](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-30-1-1024x828.png\")
<\/a><\/figure>\n\n\n\nThen update it with your authentication details. You can get the credentials from the .env<\/strong> file you created earlier.<\/p>\n\n\n\n
{
“mcpServers”: {
“mcp-atlassian”: {
“command”: “node”,
“args”: [“\/path\/to\/jira-mcp-server\/index.js”],
“env”: {
“JIRA_HOST”: “your-domain.atlassian.net”,
“JIRA_EMAIL”: “you@example.com”,
“JIRA_API_TOKEN”: “your-api-token”
}
}
}
}<\/p>\n\n\n\nRestart the Claude Desktop, and you\u2019ll see confirmation that \u201cAtlassian Jira\u201d was successfully connected.<\/p>\n\n\n\n
Output:<\/p>\n\n\n\n
Connected tool: Atlassian Jira (jira-mcp)<\/p>\n\n\n\n
![Connected tool - Atlassian Jira](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-26-1.png\")
<\/a><\/figure>\n\n\n\nStep 8: Test the Integration<\/h3>\n\n\n\n
Now you can use natural language in Claude.<\/p>\n\n\n\n
![Use natural language in Claude](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-27-1.png\")
<\/a><\/figure>\n\n\n\n![Run a command in your Claude chat input](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-29-1.png\")
<\/a><\/figure>\n\n\n\nRun this command in your Claude chat input:<\/p>\n\n\n\n
\u201cCreate a Jira issue titled \u2018Fix login bug in API\u2019<\/em> in the project DEV<\/em>.\u201d<\/p>\n\n\n\n
Claude Output:<\/p>\n\n\n\n
\u2705 Issue DEV-112 created successfully<\/p>\n\n\n\n
You can also try JQL queries like:<\/p>\n\n\n\n
\u201cShow all issues assigned to me in project DEV.\u201d<\/p>\n\n\n\n
Output:<\/p>\n\n\n\n
[
{ “key”: “DEV-12”, “summary”: “Update login API”, “status”: “In Progress”, “assignee”: “You” },
{ “key”: “DEV-15”, “summary”: “Fix UI layout bug”, “status”: “To Do”, “assignee”: “You” }
]<\/p>\n\n\n\nStep 9: Integrate with Workato Agentic MCP<\/h3>\n\n\n\n
If you\u2019d rather skip the manual setup and focus on building intelligent workflows, Workato\u2019s Agentic MCP<\/a> provides a secure, enterprise-ready alternative<\/a>.<\/p>\n\n\n\n
Workato offers a pre-spun, fully managed MCP environment with native Jira support already configured. This means you don\u2019t need to deploy or maintain your own server. You can connect your AI agents or automation pipelines to Jira instantly through Workato\u2019s MCP infrastructure.<\/p>\n\n\n\n
Best Practices<\/h2>\n\n\n\n
Use these best practices to maintain a secure, reliable, and scalable Jira MCP environment.<\/p>\n\n\n\n
- \n
- Separate environments for dev, staging, and production MCP servers.<\/li>\n\n\n\n
- Use containerization for consistent builds.<\/li>\n\n\n\n
- Log all requests and responses for auditing.<\/li>\n\n\n\n
- Apply caching for repetitive JQL queries.<\/li>\n\n\n\n
- Adopt Workato\u2019s pre-spun Jira MCP instance<\/a> to save setup time and ensure enterprise-grade compliance. Explore the Workato Enterprise MCP guide<\/a> to get started.<\/li>\n<\/ol>\n\n\n\n
![Integrating Jira with MCP transforms how development and operations teams work](\"https:\/\/blog2workato.wpengine.com\/wp-content\/uploads\/2026\/10\/image-28-1.png\")
<\/a><\/figure>\n\n\n\nConclusion<\/h2>\n\n\n\n
Integrating Jira with MCP transforms how development and operations teams work. It bridges human collaboration in Jira with AI-driven execution and automation, enabling faster, smarter workflows.<\/p>\n\n\n\n
By following this setup guide, you\u2019ve learned how to:<\/p>\n\n\n\n
- \n
- Build a Jira MCP server with Node.js<\/a>,<\/li>\n\n\n\n
- Configure Claude Desktop as your connected AI Agent, and<\/li>\n\n\n\n
- Test the integration.<\/li>\n<\/ul>\n\n\n\n
However, building and securing an MCP server from scratch can be complex. <\/p>\n\n\n\n
That\u2019s why you need to adopt Workato\u2019s secure, pre-spun Jira MCP instance, like Workato\u2019s Enterprise MCP<\/a>. Try it today and accelerate your agentic automation journey instantly!<\/p>\n\n\n\n
- Build a Jira MCP server with Node.js<\/a>,<\/li>\n\n\n\n
- Windows: %APPDATA%Claudeclaude_desktop_config.json<\/strong><\/li>\n<\/ol>\n\n\n\n
- The createIssue<\/strong> tool allows you to add new issues from Claude or any MCP client.<\/li>\n\n\n\n
- jira.js<\/strong>: official Jira client supporting REST API v3.<\/li>\n\n\n\n
- Click \u201cCreate API token\u201d<\/strong><\/li>\n\n\n\n
- Visit: https:\/\/id.atlassian.com\/manage-profile\/security\/api-tokens<\/a><\/li>\n\n\n\n
- Hybrid setup<\/li>\n<\/ol>\n\n\n\n
- Dockerized MCP<\/li>\n<\/ol>\n\n\n\n
- Building your own self-hosted MCP server (Node.js example in this guide)<\/li>\n<\/ol>\n\n\n\n
- Using Workato\u2019s pre-spun Jira MCP instance<\/li>\n<\/ol>\n\n\n\n
- Execution<\/li>\n<\/ol>\n\n\n\n
- Translation to API Call<\/li>\n<\/ol>\n\n\n\n
- Interpretation and Validation<\/li>\n<\/ol>\n\n\n\n
- Request Initiation<\/li>\n<\/ol>\n\n\n\n
- Connection and Authentication<\/li>\n<\/ol>\n\n\n\n