Confluence to Markdown MCP Server¶

MCP server for hybrid search over locally-saved and indexed Confluence content. Sync your Confluence spaces to local markdown files, then search everything from any AI agent - no Confluence API rate limits, offline-capable.

What It Does¶

Confluence Sync - Export Confluence spaces to local markdown files
Indexed Search - Full-text search across all synced Confluence content
Hybrid Mode - Search local index first, fall back to Confluence API for live data
Offline-First - Cached content available even when Confluence is down

Business Operator Use Cases¶

Use Case	Value
Knowledge Base Search	"Find the Q2 2026 product roadmap page" → instant result from local index
Cross-Space Research	"Show me all pages mentioning 'pricing change' across all spaces"
Offline Documentation	Access Confluence docs when VPN is down or traveling
AI Agent Context	Feed Confluence knowledge into agent prompts without hitting API limits

Installation¶

git clone https://github.com/gkoos/confluence2md-mcp
cd confluence2md-mcp
pip install -r requirements.txt

Configuration¶

Add to your MCP client config:

{
  "mcpServers": {
    "confluence2md": {
      "command": "python",
      "args": ["path/to/confluence2md-mcp/server.py"],
      "env": {
        "CONFLUENCE_URL": "https://your-instance.atlassian.net/wiki",
        "CONFLUENCE_USERNAME": "your-email@company.com",
        "CONFLUENCE_API_TOKEN": "your-api-token",
        "LOCAL_INDEX_PATH": "/path/to/confluence-cache"
      }
    }
  }
}

Auth¶

Atlassian API token authentication. Generate an API token at https://id.atlassian.com/manage-profile/security/api-tokens.

Tools Provided¶

sync_space - Export a Confluence space to local markdown
search_content - Full-text search across synced content
get_page - Retrieve a specific page by title or ID
refresh_index - Update local index with recent Confluence changes
list_spaces - View all synced spaces

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations¶

0 stars, brand new - Created July 9, 2026. Early development.
Initial sync time - First sync of large Confluence instances can be slow (1000s of pages).
No real-time updates - Content is only as fresh as your last sync.
Confluence license required - You need access to a Confluence instance.
Not official Atlassian - Community project, not Atlassian-supported.

Operator Verdict¶

★★★ Essential for Confluence-heavy organizations. This solves the AI agent knowledge gap for teams that live in Confluence. The offline-first approach is smart for large instances where Confluence API rate limits are a real bottleneck. Pair with the Atlassian MCP server for live Jira+Confluence access.