execute_cql_search
Execute a CQL query on Confluence to search pages
LABSCO SUMMARY
CQL is the reason this beats keyword search.
Exposing Confluence's query language rather than a simplified search box means an assistant can ask for pages in one space, modified since a date, by a given author — which is how you find the right document in a wiki that has thousands. Writing is included, and it is a whole-body replacement, so treat update_page_content as the last step of a read-edit-write sequence rather than a way to add a line.
What it is
A small MCP server for Confluence Cloud. Search runs through CQL, Confluence's own query language, so the filtering happens on the server rather than by pulling pages down and reading them; content can then be fetched by page id and written back.
What you get
- Pages found with a CQL query, with a result limit that defaults to 10 —
execute_cql_search - The content of a page fetched by its page id —
get_page_content - A page's content replaced, optionally with a new title —
update_page_content
Requirements
An Atlassian account with access to the wiki, plus an API token created from the Atlassian security settings. Three values are needed: CONFLUENCE_URL for the wiki base URL, CONFLUENCE_API_MAIL for the account email, and CONFLUENCE_API_KEY for the token. The npm package name is jira_communication_server (0.2.0) and it is marked private, so build from a clone with npm install and npm run build. update_page_content replaces a page's body, so read the current content first if you mean to append rather than overwrite.
Setup effort
Build from source
USE CASES
01
Finding the page that actually answers a question
Run a CQL query scoped to a space and a date range instead of a keyword search across the whole wiki, then read the top result.
02
Bringing a stale document up to date
Read the current page, work out the changes in the conversation, and write the revised body back with the same or a new title.
03
Summarising documentation you did not write
Search for the related pages, pull their content, and have the model produce the overview that the wiki never had.
04
Auditing which pages have gone stale
Query by last-modified date in CQL to list the documents nobody has touched, and open the ones that matter.
3 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
get_page_content
Get the content of a Confluence page
update_page_content
Update the content of a Confluence page
BEFORE YOU INSTALL
What to know first
Tools you get
3 tools when we connected.
Requires
- node on PATH
- CONFLUENCE_URL in your environment
- CONFLUENCE_API_KEY in your environment
- CONFLUENCE_API_MAIL in your environment
CONTEXT COST
Light Standing cost
Rated by the size of the 3 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
git clone https://github.com/KS-GEN-AI/confluence-mcp-server
cd confluence-mcp-server
npm install
npm run buildMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"Confluence communication server": {
"command": "node",
"args": [
"<YOUR_PATH_TO_PROJECT>/build/index.js"
],
"env": {
"CONFLUENCE_URL": "https://<YOUR_SUBDOMAIN>.atlassian.net/wiki",
"CONFLUENCE_API_MAIL": "<YOUR_CONFLUENCE_API_MAIL>",
"CONFLUENCE_API_KEY": "KEY_FROM: https://id.atlassian.com/manage-profile/security/api-tokens"
}
}
}
}
