Resume Forge MCP
Allows browsing and fetching LaTeX resume templates from Overleaf's gallery, dynamically importing over 350 templates.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Resume Forge MCP Generate my resume using the modern template"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Resume Forge MCP
An MCP (Model Context Protocol) server for intelligent LaTeX resume generation with multiple template styles, quality scoring, and job description tailoring. Works with any MCP-compatible AI assistant — Claude Desktop, Kiro, Cursor, Windsurf, etc.
Features
3 built-in templates: Modern (color accents), Classic (traditional), Minimal (no-frills)
Overleaf integration: Browse and fetch 350+ templates dynamically
Multi-format import: Import resume data from
.tex,.pdf, or.docxfilesQuality scoring: Bullet analysis, ATS compatibility, keyword matching
Job tailoring: Auto-select and rank content based on job descriptions
Visual preview: Compile to PDF and preview as images directly in your AI assistant
Related MCP server: resume-processor-mcp
Quick Start
1. Import my resume from resume.pdf
2. List available templates
3. Generate my resume using the modern template
4. Score my resume against keywords: Python, AWS, KubernetesThat's it — 4 prompts to go from a PDF to a scored, template-styled resume.
Installation
uvx resume-forge-mcpWith PDF and DOCX import support:
uvx --with pymupdf --with python-docx resume-forge-mcpPrerequisites
For PDF compilation (compile_and_preview), you need LaTeX installed:
macOS:
brew install --cask mactexUbuntu:
sudo apt install texlive-latex-base texlive-latex-extra texlive-fonts-extraWindows: Install MiKTeX
LaTeX is only needed for compiling to PDF. All other tools (import, generate .tex, score, tailor) work without it.
Configuration
Add to your AI assistant's MCP config:
Claude Desktop
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"resume-forge": {
"command": "uvx",
"args": ["resume-forge-mcp"]
}
}
}Kiro
Add to ~/.kiro/settings/mcp.json:
{
"mcpServers": {
"resume-forge": {
"command": "uvx",
"args": ["resume-forge-mcp"]
}
}
}Custom directories (optional)
{
"mcpServers": {
"resume-forge": {
"command": "uvx",
"args": ["resume-forge-mcp"],
"env": {
"RESUME_DATA_DIR": "/path/to/data",
"RESUME_TEMPLATE_DIR": "/path/to/templates",
"RESUME_OUTPUT_DIR": "/path/to/output"
}
}
}
}Environment Variables
Variable | Default | Purpose |
|
| Imported resume data (JSON) |
|
| Fetched Overleaf templates |
|
| Generated |
All env vars are optional — defaults to ~/.resume-forge/ subdirectories on any OS.
Tools (11)
Templates — Pick a style or browse Overleaf
Tool | What it does |
| Shows 3 built-in templates (modern, classic, minimal) and tells you Overleaf is available for 350+ more |
| Browse Overleaf's gallery by category (cv, cover-letter, etc.) — returns names and URLs |
| Download a specific Overleaf template by URL, optionally save it locally |
Data — Get your resume in and edit it
Tool | What it does |
| Feed it a |
| View everything that was imported — all sections, all entries, all bullets |
| Fix a company name, add a bullet, delete an old job, update your phone number — works on any section |
Generate — Create your resume
Tool | What it does |
| Takes your data + a template (modern/classic/minimal) → renders a complete |
| Compiles |
| Paste a job description → it parses requirements, picks your most relevant experience, generates a tailored resume |
Analyze — Improve your resume
Tool | What it does |
| Scores every bullet (action verb? metrics? right length?), checks ATS compatibility, matches keywords you provide |
| Shows your directories, pdflatex status, and all available tools |
Usage Examples
Import and generate
> Import my resume from /path/to/resume.pdf
> Generate my resume using the classic templateScore and improve
> Score my resume quality against keywords: Java, AWS, Kafka, microservices
> Update experience index 0 bullet 2 to add a metricTailor for a job
> Generate a tailored resume for this job description:
> Senior Backend Engineer - 5+ years Java, AWS, event-driven architecture...Browse Overleaf templates
> Browse Overleaf templates for CVs
> Fetch the template at https://www.overleaf.com/latex/templates/...How It Works
Import — Your resume file is parsed into structured JSON (contact, experience, education, skills, publications)
Edit — Fix any parsing issues or add new content via
update_resume_dataTemplate — Pick a built-in style or fetch one from Overleaf
Generate — Your data is rendered through a Jinja2 template into a
.texfileCompile — pdflatex compiles the
.texto a PDF (optional — requires LaTeX installed)Tailor — Paste a job description and get a resume with auto-selected relevant content
License
MIT
Available Tools
11 toolsbrowse_overleaf_templatesA
Browse Overleaf's template gallery to discover LaTeX templates.
Args: tag: Gallery category (cv, cover-letter, report, presentation, etc.) page: Page number (1-indexed).
Returns: JSON with list of templates (name, url) and pagination info.
Inputs: tag, page. Returns: result
compile_and_previewA
Compile a LaTeX resume to PDF and return a preview image.
If tex_path is provided, compiles that file directly. Otherwise generates from resume data using the specified template, then compiles.
Args: tex_path: Path to existing .tex file. If None, generates from data first. output_filename: Base filename for output files (used when generating). template_name: Template style when generating ('modern', 'classic', 'minimal'). dpi: Resolution for preview image (default 200).
Returns: Image of the rendered resume, or error JSON string.
Inputs: tex_path, output_filename, template_name, dpi.
fetch_overleaf_templateA
Fetch LaTeX source code from an Overleaf template page.
Args: template_url: Full Overleaf template URL. save_as: Optional filename to save locally (without .tex extension).
Returns: JSON with template name, source code (or saved path), and status.
Inputs: template_url (required), save_as. Returns: result
generate_resumeB
Render resume data to LaTeX source using a template.
Args: output_filename: Base filename for output (without extension). template_name: Template style ('modern', 'classic', 'minimal').
Returns: JSON with LaTeX file path and status.
Inputs: output_filename, template_name. Returns: result
generate_tailored_resumeA
Tailor resume to a job description: parse JD, select content, generate.
Parses the job description, selects and ranks experiences/projects by relevance, generates a tailored resume, and optionally compiles to PDF.
Args: jd_text: Raw job description text. template_name: Template style ('modern', 'classic', 'minimal'). output_filename: Base filename for output. target_tags: Optional tags to prioritize (e.g., ["swe", "ml"]). include_experiences: Force include these experience indices. exclude_experiences: Exclude these experience indices. include_projects: Force include these project indices. exclude_projects: Exclude these project indices. max_experiences: Maximum experience entries (default 4). max_projects: Maximum project entries (default 3). compile_pdf: Whether to compile and preview (default True). dpi: Resolution for preview image.
Returns: Preview image if compile_pdf=True, otherwise JSON with details.
Inputs: jd_text (required), template_name, output_filename, target_tags, +8 more.
get_configA
Show current configuration and tool availability.
Returns: JSON with configuration details.
Inputs: no named parameters. Returns: result
get_resume_dataA
Read the master resume data pool.
Returns the complete resume data including all entries across all sections.
Returns: JSON with complete resume data or error if not found.
Inputs: no named parameters. Returns: result
import_resumeA
Import resume data from a file (PDF, DOCX, or LaTeX .tex).
Parses the file and extracts contact, education, experience, projects, and skills into structured data. The imported data becomes the master resume pool used by generate_resume.
Args: file_path: Absolute path to the file (.tex, .pdf, or .docx).
Returns: JSON with import status and summary of extracted entries.
Inputs: file_path (required). Returns: result
list_templatesA
List available resume templates.
Shows built-in templates for generate_resume, plus info on fetching additional templates from Overleaf dynamically.
Returns: JSON with built-in templates and Overleaf integration details.
Inputs: no named parameters. Returns: result
score_resume_qualityA
Analyze resume quality with detailed scoring.
Evaluates bullet quality, ATS compatibility, keyword matching, and page layout. Returns actionable improvement suggestions.
Args: keywords: Optional list of keywords to match (from job description).
Returns: JSON with detailed quality scores and suggestions.
Inputs: keywords. Returns: result
update_resume_dataA
Add, edit, or remove entries in the master resume data.
Args: section: Section to modify (contact, experience, projects, education, publications, skills). action: Action to perform (add, update, delete). Contact only supports update. index: Index of entry to update/delete (required for update/delete, not used for contact). data: JSON string or dict of entry data (required for add/update).
Returns: JSON with update status.
Inputs: section (required), action (required), index, data. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 11 tools
Tools are mostly distinct, but generate_resume and generate_tailored_resume have similar names and overlapping functionality, which could cause confusion. Other tools are clearly separated.
All tools use consistent verb_noun snake_case pattern. Even compile_and_preview, though using 'and', follows a predictable verb_verb structure.
With 11 tools, the server covers the full resume workflow without being bloated. Each tool serves a clear purpose.
The toolset covers import, data management, template discovery, generation, tailoring, compilation, and quality scoring. A minor gap is that compile_and_preview only returns an image, not a direct PDF download.
Maintenance
Related MCP Connectors
Resume builder with native MCP — create and edit resumes from your AI assistant.
Free open-source resume builder with remote MCP tools for resumes and job applications.
MCP server for VC pitch-deck scoring, thesis-fit matching, and deal-flow management.
MCP server for generating rough-draft project plans from natural-language prompts.
Related MCP Servers
- FlicenseAqualityDmaintenanceAn MCP server that tailors LaTeX resumes to job descriptions using Claude, returning a structured diff with keywords, gap summary, bullet changes, and a guardrails report.2-
- AlicenseNot gradedqualityDmaintenanceMCP server for processing Markdown resumes into multiple formats (PDF, LinkedIn templates) with AI analysis and Claude integration.MIT
- FlicenseNot gradedqualityDmaintenanceMCP server for AI-native resume optimization, providing tools to load JD, analyze match, rewrite sections, and assemble customized resumes.-
- AlicenseAqualityDmaintenanceMCP server for AI-powered resume optimization and generation. It enables clients to optimize existing resumes or generate new ones from background info via LLM, producing formatted .docx and .pdf files.21MIT