mcp-jira-stdio — Claude Code MCP server: Jira issues from chat
Hooks Jira into your coding agent: open tickets, run JQL searches, post comments and file new issues without leaving the chat.
claude mcp add --env JIRA_BASE_URL=<JIRA_BASE_URL> --env JIRA_EMAIL=<JIRA_EMAIL> --env JIRA_API_TOKEN=<JIRA_API_TOKEN> --env MCP_REDACT_SENSITIVE=<MCP_REDACT_SENSITIVE> --env LOG_LEVEL=<LOG_LEVEL> --transport stdio mcp-jira-stdio -- npx -y mcp-jira-stdioWhat mcp-jira-stdio does
The server exposes Jira as a set of agent tools. jira_get_issue pulls a ticket by key, jira_search_issues runs JQL with paging and field selection, jira_create_issue and jira_update_issue write changes back. Subtasks, comments and reference data come along too: issue types, priorities, statuses, custom fields. Before creating anything, jira_get_create_meta lists the fields a project demands, so the agent hits fewer rejections on required custom fields.
It speaks stdio and starts from npm with npx -y mcp-jira-stdio. You need Node.js 20 or newer plus three values: the Jira instance URL, your account email and an API token, passed as JIRA_BASE_URL, JIRA_EMAIL and JIRA_API_TOKEN. Two more variables, MCP_REDACT_SENSITIVE and LOG_LEVEL, control masking of sensitive values in logs and log verbosity.
Grab the token in Atlassian's account security settings. With a stale or wrong token the API rejects every call and you get auth errors instead of tickets, which is the first thing to check when the setup looks fine but nothing loads.
What it helps with
- open a Jira ticket by key without a browser tab
- pull issues by JQL and sort them by status
- file a new issue with the fields a project requires
- drop a comment on a ticket from the agent chat
- see which fields are mandatory before creating a task
Example prompts for the agent
- Show me everything on
PROJ-142, description and comments included - Run a JQL search for my open bugs in
PROJand group them by priority - Create a subtask under
PROJ-142listing the reproduction steps
How to install mcp-jira-stdio
Platform CLI
claude mcp add --env JIRA_BASE_URL=<JIRA_BASE_URL> --env JIRA_EMAIL=<JIRA_EMAIL> --env JIRA_API_TOKEN=<JIRA_API_TOKEN> --env MCP_REDACT_SENSITIVE=<MCP_REDACT_SENSITIVE> --env LOG_LEVEL=<LOG_LEVEL> --transport stdio mcp-jira-stdio -- npx -y mcp-jira-stdioReplace <…> with your own values
Platform CLI
codex mcp add mcp-jira-stdio --env JIRA_BASE_URL=<JIRA_BASE_URL> --env JIRA_EMAIL=<JIRA_EMAIL> --env JIRA_API_TOKEN=<JIRA_API_TOKEN> --env MCP_REDACT_SENSITIVE=<MCP_REDACT_SENSITIVE> --env LOG_LEVEL=<LOG_LEVEL> -- npx -y mcp-jira-stdioReplace <…> with your own values
Config file
{
"mcpServers": {
"mcp-jira-stdio": {
"command": "npx",
"args": [
"-y",
"mcp-jira-stdio"
],
"env": {
"JIRA_BASE_URL": "<JIRA_BASE_URL>",
"JIRA_EMAIL": "<JIRA_EMAIL>",
"JIRA_API_TOKEN": "<JIRA_API_TOKEN>",
"MCP_REDACT_SENSITIVE": "<MCP_REDACT_SENSITIVE>",
"LOG_LEVEL": "<LOG_LEVEL>"
}
}
}
}Replace <…> with your own values
Config file
{
"servers": {
"mcp-jira-stdio": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"mcp-jira-stdio"
],
"env": {
"JIRA_BASE_URL": "<JIRA_BASE_URL>",
"JIRA_EMAIL": "<JIRA_EMAIL>",
"JIRA_API_TOKEN": "<JIRA_API_TOKEN>",
"MCP_REDACT_SENSITIVE": "<MCP_REDACT_SENSITIVE>",
"LOG_LEVEL": "<LOG_LEVEL>"
}
}
}
}Replace <…> with your own values
Platform CLI
gemini mcp add -e JIRA_BASE_URL=<JIRA_BASE_URL> -e JIRA_EMAIL=<JIRA_EMAIL> -e JIRA_API_TOKEN=<JIRA_API_TOKEN> -e MCP_REDACT_SENSITIVE=<MCP_REDACT_SENSITIVE> -e LOG_LEVEL=<LOG_LEVEL> mcp-jira-stdio npx -y mcp-jira-stdioReplace <…> with your own values
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about mcp-jira-stdio
How do I install mcp-jira-stdio in Claude Code?
Run `claude mcp add mcp-jira-stdio -e JIRA_BASE_URL=<JIRA_BASE_URL> -e JIRA_EMAIL=<JIRA_EMAIL> -e JIRA_API_TOKEN=<JIRA_API_TOKEN> -- npx -y mcp-jira-stdio`, replacing the placeholders with your own credentials.
What does mcp-jira-stdio do?
It wraps the Jira API as MCP tools. The agent can read and edit issues, search with JQL, add comments and subtasks, and look up project metadata.
What does it need to run?
Node.js 20 or newer, and a Jira API token along with the instance URL and account email. The token is issued in Atlassian's security settings.
