Skip to main content
Glama
Sign In
confluence-mini-mcp

confluence-mini-mcp

by bonkey
OverviewSchemaSourceRelated ServersDiscussionsAdmin
Python
Remote

confluence-mini-mcp

Lightweight MCP server that crawls Confluence subtrees and serves cached Markdown. Works with Claude Code and nanobot (Slack) via stdio.

Claude Code config

{
  "mcpServers": {
    "confluence-docs": {
      "command": "uvx",
      "args": ["--from", "git+https://github.com/bonkey/confluence-mini-mcp", "confluence-mini-mcp"],
      "env": {
        "CONFLUENCE_BASE_URL": "https://company.atlassian.net/wiki",
        "CONFLUENCE_EMAIL": "you@company.com",
        "CONFLUENCE_API_TOKEN": "...",
        "CONFLUENCE_ROOT_PAGE_IDS": "123456,789012"
      }
    }
  }
}

Pin to a tag: "--from", "git+https://github.com/bonkey/confluence-mini-mcp@v0.1.1"

Related MCP server: confluence-mcp-server

Nanobot config

[mcp.confluence]
command = "uvx"
args = ["--from", "git+https://github.com/bonkey/confluence-mini-mcp", "confluence-mini-mcp"]
env = { CONFLUENCE_BASE_URL = "https://company.atlassian.net/wiki", CONFLUENCE_EMAIL = "you@company.com", CONFLUENCE_API_TOKEN = "...", CONFLUENCE_ROOT_PAGE_IDS = "123456,789012" }

Tools

Tool

Description

search_pages

Keyword search across cached pages (title 10x weight)

get_page

Full Markdown content by page ID (Confluence or external)

list_pages

Index of all cached pages (optional space_key filter)

refresh_cache

Force re-crawl, bypass TTL

Crawling

The crawler walks a graph, not just a tree:

  • Child pages — standard parent → children traversal

  • Linked Confluence pages — any internal link in page content is followed (enables a single "hub page" linking to multiple subtrees)

  • External HTTP links — fetched as Markdown (1 level deep, follows redirects only)

In-page directives

Add a [confluence-mcp] block anywhere on a Confluence page to control crawl behaviour for that subtree:

[confluence-mcp]
max_depth=3
follow_links=false
follow_external=true

Directive

Default

Description

max_depth

(global config)

Override crawl depth from this page down

follow_links

true

Follow linked Confluence pages in content

follow_external

true

Fetch external HTTP links in content

The directive block is stripped from the Markdown output — agents never see it.

Configuration

All settings via env vars or confluence-mini-mcp.toml:

Env var

Description

Default

CONFLUENCE_BASE_URL

Wiki base URL

required

CONFLUENCE_EMAIL

Atlassian email

required

CONFLUENCE_API_TOKEN

API token

required

CONFLUENCE_ROOT_PAGE_IDS

Comma-separated root page IDs

required

CONFLUENCE_CACHE_DIR

Cache directory

~/.cache/confluence-subtree-mcp

CONFLUENCE_CACHE_TTL_MINUTES

Cache max age

30

CONFLUENCE_MAX_DEPTH

Max crawl depth

10

CONFLUENCE_MAX_PAGES

Max pages to crawl

500

CONFLUENCE_REFRESH_INTERVAL

Background refresh interval (0=off)

5

CONFLUENCE_DRY_RUN

Fake cache mode, no API calls

false

Development

just dev     # run in dry-run mode
just check   # format + test
just release # tag + push

All tasks in the Justfile. Requires uv and just.

Available Tools

4 tools
get_pageA

Retrieve the full Markdown content of a specific page by ID.

Use this after search_pages to read the complete content of a result. Pages may be Confluence pages or crawled external websites.

Args: page_id: The page ID from search_pages or list_pages results.

Inputs: page_id (required). Returns: object

list_pagesB

List all cached pages, optionally filtered by space key.

Args: space_key: Optional space key to filter by.

Inputs: space_key. Returns: result

refresh_cacheA

Force an immediate re-crawl of all configured subtrees, bypassing TTL and version checks.

Inputs: no named parameters. Returns: object

search_pagesA

Full-text search across all cached Confluence pages and external links.

Returns short snippets only. IMPORTANT: Always call get_page on relevant results to read the full content before answering questions. The snippet is just a preview — the full page often contains the information you need.

Args: query: Space-separated keywords to search for. Try single key terms first; the search supports prefix matching. limit: Max results to return (default 10, max 20).

Inputs: query (required), limit. Returns: result

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Deploy Server
    One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

    TDQS

    A4/5.0

    Scored across 4 tools

    Disambiguation4/5

    list_pages and search_pages both surface pages, but they differ by enumeration vs full-text query and are clearly documented. get_page and refresh_cache have non-overlapping cache-read and cache-maintenance roles, so misselection is unlikely.

    Naming Consistency5/5

    All four tools follow a consistent snake_case verb_noun pattern: list_pages, refresh_cache, search_pages, get_page. There is no mixing of naming conventions.

    Tool Count5/5

    Four tools are well-scoped for a mini read-oriented Confluence cache MCP. Each tool has a distinct role: enumerate, search, retrieve, and refresh, with no redundant tools.

    Completeness4/5

    The set covers the core read path: list cached pages, search them, and fetch full page content. It lacks page creation/update/delete, but as a mini cache reader this is a reasonable scope with only minor gaps.

    Resources

    Related MCP Connectors

    Related MCP Servers