All Guides

Rezi MCP Server

Updated on
April 10, 2026
Rezi MCP Server
Contributor
Jacob Jacquet
April 10, 2026

Rezi MCP Server

The Rezi Model Context Protocol (MCP) server lets AI tools connect to a user's Rezi account and work directly with resume data. Instead of copying resume JSON into chat, users can access their resumes, read them, and write updates back to Rezi through MCP.

Who can use this feature

  • Any Rezi user with a Rezi Pro subscription
  • Any MCP client that supports remote streamable HTTP servers and interactive tool use
  • Examples: Codex, Claude Code, Claude Desktop, Cursor, Gemini CLI, Lovable, and other MCP clients with remote HTTP support

Setup instructions

Codex‍

codex mcp add rezi --url https://api.rezi.ai/mcp

Claude Code‍

claude mcp add rezi --transport http https://api.rezi.ai/mcp

Claude Desktop

Open Claude Desktop > Customize > Connectors > Add custom connector with URL https://api.rezi.ai/mcp. Do not add through claude_desktop_config.json.

Cursor‍

Add to mcp.json: { "mcpServers": { "rezi": { "url": "https://api.rezi.ai/mcp", "transport": "streamable-http" } } }

Gemini CLI

gemini mcp add --transport http rezi https://api.rezi.ai/mcp

Or add manually in ~/.gemini/settings.json: { "mcpServers": { "rezi": { "httpUrl": "https://api.rezi.ai/mcp" } } }

Lovable

Settings > Connectors > Personal connectors > New MCP server. Set URL to https://api.rezi.ai/mcp.

Any streamable HTTP MCP client‍

Point at https://api.rezi.ai/mcp

Available tools

  • list_resumes: [regular] Returns the authenticated user's resumes as lightweight summaries with ID, name, job title, and last updated time
  • read_resume: [regular] Returns the full JSON document for a single resume by resume_id
  • write_resume: [regular] Creates a new resume when resume_id is omitted, or updates an existing resume when resume_id is provided

Resume write shape

write_resume expects a resume object with top-level fields: name, jobTitle, jobDescription, jobCompany, template, data.

The data object includes sections: contact, summary, experience, education, skills, projects, certification, coursework, involvement, awards. Collection-style sections use UUID keys, plus index for ordering and hide for visibility.

How resume access works

  • list_resumes returns compact summaries for the authenticated user
  • read_resume returns the complete stored resume document
  • write_resume writes back through the authenticated user context (create or update)
  • On updates, Rezi MCP reads the existing resume and verifies ownership before writing

Session behavior and security

  • Credentials stored in memory, scoped to MCP session ID
  • Session credentials refreshed automatically near expiry
  • Reads/writes use the authenticated user's Rezi session context
  • No credentials persisted to disk. Service restart requires re-sign-in
  • Access aligned with user's own Rezi permissions

Server URL

https://api.rezi.ai/mcp