MCP Documentation

Remnus MCP is a Model Context Protocol server built into every Remnus workspace. It lets AI agents — Claude, Cursor, Windsurf, and any MCP-compatible client — read and write your workspace data over a standard HTTP API.

What you can do

• Query pages and databases with full-text search and property filters
• Create and update pages, database rows, and entire databases
• Automate reports, task extraction, and kanban triage with built-in prompts
• Monitor all agent activity through a structured audit log
• Prepare compact, OKF-aware task context and optionally require it before MCP writes

Endpoint

https://www.remnus.com/api/mcp

Always use the www host — the apex remnus.com redirects to www.remnus.com, and some OAuth clients reject the resulting resource-indicator mismatch. Uses the modern Streamable HTTP transport (stateless, one request per call) — the transport the current MCP spec recommends for remote servers.

Quick start

1. Open your workspace → sidebar AI Agents button → Connect editor
2. Pick your editor. Most connect with OAuth — no token to copy, just approve the consent screen in your browser on first connect. Prefer a token? Expand Advanced to mint one.
3. Ask your agent to prepare a read-only context pack — see Getting Started for the full walkthrough

Documentation

Getting Started	OAuth connect, PAT fallback, first call
Connect Your Editor	Windsurf, Continue, Antigravity, Cline, Zed & more — ready configs + OAuth
Authentication	Bearer tokens, scopes, rate limits
Read Tools	10 read-only tools
Write Tools	10 write tools
Resources	6 MCP resource templates
Prompts	7 built-in prompt templates
Agent Memory	Save & recall durable memory over MCP
Token-Efficient Usage	Cut a typical read 80–90% with projection, outline, digest & delta
Context-first MCP	Smart/Strict automation, contextRunId, trust and client boundaries