MCP overview

April 2, 2025

Krafter exposes a hosted MCP (Model Context Protocol) server, so you can use any MCP-compatible AI assistant to create and edit resumes without leaving your AI tool.

What is MCP?

Model Context Protocol is an open standard that lets AI assistants call server-side tools. Krafter's MCP server exposes resume operations as tools — meaning you can ask Claude or Cursor to "create a resume for Jane Doe" and it will call Krafter's API on your behalf.

What you can do via MCP

List all your resumes
Create a new resume from natural-language input
Read the full content of any resume
Update an entire resume or just a single section
Delete or duplicate a resume
Browse available templates, accent colors, and density presets
Update visual settings — template, font, accent color, density, spacing
Get the resume JSON schema so the AI understands the data format

Available tools

Krafter’s MCP server provides 10 tools:

get_resume_schema — JSON Schema + example for creating/updating resumes
list_resumes — lists all your resumes with IDs and editor URLs
get_resume — reads the full content of a resume by ID
create_resume — creates a new resume from structured data
update_resume — replaces the full content of an existing resume
delete_resume — soft-deletes a resume so it no longer appears in your list
duplicate_resume — clones a resume into a new copy
list_templates — returns all templates, accent color presets, and density presets
update_settings — changes visual settings (template, font, colors, spacing) without touching content
update_section — replaces items in a single section by type (e.g. work, skills, education)

Supported clients

Claude Desktop — full tool-use support via Streamable HTTP
Cursor — add Krafter as an MCP server in .cursor/mcp.json
Zed — supported via the MCP extension
Any other MCP-compatible client that supports Streamable HTTP transport

Getting started

1. Generate an API key in Krafter under Settings → MCP. 2. Follow the setup guide for your specific client below.