Triton AI Docs

Codex

Connect Jira and Confluence through the Codex app, CLI, or IDE extension.

First complete Set up Jira and Confluence.

Add the server

Codex uses ~/.codex/config.toml for user configuration. The app, CLI, and IDE extension share this file. On native Windows, it is %USERPROFILE%\.codex\config.toml unless you changed CODEX_HOME.

Add this table to your existing file. Replace YOUR_NAME and the executable path with your local values.

~/.codex/config.toml
[mcp_servers.atlassian]
command = "/Users/YOUR_NAME/.local/bin/uvx"
args = ["mcp-atlassian==0.23.0", "--transport", "stdio", "--env-file", "/Users/YOUR_NAME/.config/tritonai/atlassian.env"]
startup_timeout_sec = 60

The Windows example uses TOML literal strings, so backslashes stay unchanged. Keep any existing model-provider configuration and other MCP servers.

See the official Codex MCP reference for the configuration fields.

Check the connection

Restart Codex. In the CLI, run /mcp to inspect the connection and tools. You can also list the registered servers from a terminal:

codex mcp list

Check for the four tools from the shared setup. Run the Jira and Confluence checks to check authentication as well as registration.

If you use Codex inside WSL, use the WSL home directory and executable paths. See Codex on Windows.

On this page