The Deframe MCP Server is hosted and ready to use — no installation required. Point your AI host at the server URL and add your API key to start using all 25 tools and 9 prompts. Server URL:Documentation Index
Fetch the complete documentation index at: https://docs.deframe.io/llms.txt
Use this file to discover all available pages before exploring further.
http://mcp.deframe.io/mcp
Prerequisites
- A Deframe API key — get one here
- An MCP-compatible host: Claude Desktop or similar
- Node.js installed (required for
npx)
Claude Desktop
Open the configuration file
- macOS
- Windows
Open
~/Library/Application Support/Claude/claude_desktop_config.jsonAdd the server entry
Claude Desktop uses
mcp-remote as a bridge to connect to remote HTTP servers. The --allow-http flag is required because mcp-remote enforces HTTPS by default.Claude Code
Add the server
- CLI
- Config file
Claude Code supports HTTP transport natively — no
mcp-remote required.Cursor
Add the server
- Config file
- UI
Create or edit
.cursor/mcp.json (project) or ~/.cursor/mcp.json (global):The
--allow-http flag is required because mcp-remote enforces HTTPS by default.Google Gemini CLI
OpenAI Codex
Other Agents
Any MCP-compatible host that accepts amcpServers JSON config can connect to the Deframe server using mcp-remote:
Replace
your-api-key with your Deframe API key. The --allow-http flag is required because mcp-remote enforces HTTPS by default.Next Steps
Tools Reference
Browse all 25 tools and 9 prompts with inputs and outputs
MCP Server Overview
Learn what the server does and when to use it