LangFlow-MCP-High-ATS-Resume-creator

# 🧠 LangFlow MCP High ATS Resume Creator (ATS-Aware) This repository contains a LangFlow-exported `.json` file representing a **custom multi-agent flow**, designed to act as an **MCP server**. This flow can be used with **Claude Desktop** or **any MCP-compatible client**. Its core functionality is to **generate a high ATS (Applicant Tracking System) score resume** based on a LinkedIn job post URL and a user's existing resume. The system parses both sources, summarizes relevant content, and produces a tailored, ATS-friendly resume in multiple formats. --- ## 📜 Architectural Decision Records (ADRs) ### ADR 001 – Why LangFlow? LangFlow allows for **modular low-code orchestration** of agentic LLM pipelines, making it ideal for our goal of chaining multiple agents with custom components and data parsers. ### ADR 002 – Why MCP Server? By exposing this flow as an MCP server: - It becomes interoperable with **Claude Desktop** and other MCP clients. - Users can interact via chat interfaces and benefit from stateful agent behavior. - Encourages **reuse and composability** across different workflows. ### ADR 003 – Agent Design and Flow #### 🎯 Agent 1: Resume Summarizer - **Input:** User's current resume (raw text or document) - **Output:** Summary of professional experience, total years, key skills, responsibilities, and education. - ✅ Focused extraction using resume-aware prompt logic. #### 🎯 Agent 2: Job Description Summarizer - **Input:** LinkedIn job post URL via chat. - **Output:** Structured summary of required qualifications. - 🛠️ Powered by a **custom LangFlow component** that parses HTML directly from the URL input. #### 🔄 Data Parsing and Conversion - Each agent's output passes through a **Message-to-Data Converter**, followed by a **combined parser** to normalize structure and prepare inputs for resume generation. #### 🎯 Agent 3: Resume Generator - **Input:** Parsed summaries from Agent 1 & Agent 2. - **Output:** ATS-optimized resume in **LaTeX** format matching job criteria with user's skills. - 🧠 Tailored resume structure focused on score improvement. #### 🛠️ Custom LangFlow Component: Format Converter - Converts LaTeX output to: - ✅ `PDF` - ✅ `DOCX` - ✅ `TXT` - Makes the resume instantly downloadable and shareable. #### 🎯 Agent 4: ATS Score Evaluator - **Input:** Generated resume and job summary. - **Output:** Final **ATS score (out of 100)** indicating how well the resume matches the job. - ✅ Helps users iterate on improvements. ### 🔀 ADR 004 – Data Pipeline ```mermaid flowchart TD FileUpload[File Upload] --> Agent1[Agent 1: Resume Summary] ChatInput[Chat Input URL] --> MCPCustomcomponent[URL HTML Fetcher] MCPCustomcomponent --> Agent2[Agent 2: Job Summary] Agent1 --> MsgToData1[Message to Data 1] Agent2 --> MsgToData2[Message to Data 2] MsgToData1 --> Merge1[Merge Resume + Job Summary] MsgToData2 --> Merge1 Merge1 --> Parse1[Parser for Resume] Parse1 --> Agent3[Agent 3: Resume Generator in latex] Agent3 --> MsgToData3[Message to Data 3] MsgToData3 --> Parse2[Parser for LaTeX] Parse2 --> LatexToPDF[Convert LaTeX to PDF/DOCX/TXT] LatexToPDF --> ResumeOut[Output Resume] LatexToPDF --> MsgToData4[Message to Data 4] Agent2 --> MsgToData5[Message to Data 5] MsgToData4 --> Merge2[Merge Resume + Job Summary for Scoring] MsgToData5 --> Merge2 Merge2 --> Parse3[Parser for Score Input] Parse3 --> Agent4[Agent 4: ATS Scoring] Agent4 --> ScoreOut[Output Score] ``` --- ## 🔗 Compatibility ✅ Works with: - Claude Desktop (as MCP Client) - Any custom MCP-compatible chat client 📦 LangFlow Version: `v1.4+` --- ## 🧪 Example Use Case > **Prompt:** > "Please generate an tailored resume for [LinkedIn Job URL]" **Result:** - Tailored resume (PDF, DOCX, TXT) - ATS match score out of 100 - Explanation of missing or weak areas --- ## 🚀 Get Started 1. Clone this repo 2. Import the `.json` into LangFlow 3. Connect with MCP client (e.g., Claude Desktop) 4. Start chatting! --- ## 🧩 Contributing Have ideas for new agents or better format converters? We welcome contributions via PR or feedback. --- ## 📄 License MIT License Copyright © 2025 ---