MCP Integration

Connect your AI assistant
to Resume-MCP

Resume-MCP implements the Model Context Protocol — giving Claude Desktop, Cursor, Zed, Windsurf, and any MCP client direct access to your resume tools.

Model Context Protocol

What is MCP?

MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude call external tools and services directly — without custom plugins or complex API wiring. You configure it once, then just ask your AI naturally.

Resume-MCP runs an MCP server at /mcp/sse using SSE transport. Point your client at that URL and every resume tool — generation, tailoring, PDF management — becomes a native capability inside your AI assistant.

No extra auth, no separate account. Works on localhost during development and on https://resume-mcp.site/mcp/sse in production.

claude_desktop_config.json
// Add to Claude Desktop settings { "mcpServers": { "resume-mcp": { "url": "https://resume-mcp.site/mcp/sse", "transport": "sse" } } } // Then ask Claude naturally: "Create a resume for a Senior DevOps Engineer with 6 years AWS experience, strong in Kubernetes and Terraform." // Or tailor to a job posting: "Tailor my resume for this JD: [paste]"

LOCAL DEVELOPMENT

localhost config
"url": "http://localhost:8000/mcp/sse"
Compatible Clients

Works with your favourite tools

Any client that speaks the Model Context Protocol can connect. Here are the most popular ones.

Claude Desktop Cursor Zed Windsurf Cline Continue.dev Any MCP client
MCP Tools

6 tools, full resume workflow

Every tool is callable from your AI assistant in plain language. No JSON wrangling required.

generate_resume_pdf

Generate PDF from scratch

Pass user details text and get back a fully compiled PDF resume. Gemini writes the LaTeX; pdflatex compiles it.

customize_resume_for_jd

Tailor to job description

Provide your resume text and a job description. The AI rewrites bullets, bolds keywords, and produces a tailored PDF.

rename_pdf

Rename a generated PDF

Rename any previously generated PDF to something more descriptive — e.g. google_swe_2025.pdf.

list_generated_pdfs

List all PDFs

Returns a list of all generated resume PDFs with filenames, sizes, and creation timestamps.

fetch_pdf

Retrieve a PDF by filename

Get the download URL for a specific PDF by its filename. Useful for sharing or re-downloading previous resumes.

delete_pdf

Delete a PDF

Permanently delete a generated PDF from the server by filename. Cleans up your resume library.

REST API

Integrate with anything

Full OpenAPI docs at /docs. Key endpoints below — all return JSON with a filename and download URL on success.

Method & Endpoint What it does
POST/api/v2/create-resume Generate a full resume from plain-text user details. AI writes LaTeX via Jinja2 template, pdflatex compiles it.
POST/api/v2/tailor-resume Tailor an existing resume to a job description. Rewrites bullets, bolds keywords, produces a new PDF.
POST/api/rename-pdf Rename a previously generated PDF to a new filename. Also renames the companion JSON if present.
GET/api/pdfs/{filename} Download a specific generated PDF by filename. Returns the file with correct Content-Type headers.
GET/api/pdfs List all generated PDF resumes with filenames, file sizes, and creation timestamps.
GET/api/stats Live counts — total generated, created from scratch, tailored to job descriptions.
GET/api/health System health check — confirms LaTeX installation and Gemini AI connectivity.
GET/mcp/sse MCP Server SSE endpoint — connect any MCP-compatible AI client here.
Full API Reference

Ready to integrate?

The interactive OpenAPI docs let you test every endpoint directly in the browser. Schemas, request bodies, response formats — all documented.

Open Full API Docs →