<h2> The Problem </h2> <p>More and more companies are using AI for resume screening. HR hands the JD to an Agent, and the Agent matches candidates. But resumes are essentially PDFs (or web pages) — hard for AI to parse cleanly. Layout noise, scattered information, no structured querying.</p> <p>So I asked myself: why not make my resume <strong>AI-native</strong>? Not "AI-friendly" marketing fluff, but something an AI Agent can actually connect to, query on demand, and run structured analysis against.</p> <p>I built exactly that: a MCP Server for my resume. Visit the public page, and an AI Agent can auto-discover the endpoint, connect, search through experience, analyze job fit, and query skill proficiency.</p> <p>The project is open source: <a href="https://github.com/Leeson-Wong/resume-maker" rel="noopener noreferrer">github.com/Leeson-Wong/resume-maker</a></p> <p>This article won't oversell the technical implementation (honestly, it's basic). Instead, let's talk about what the <strong>Resume + MCP</strong> combination actually means.</p> <h3> What is MCP? </h3> <p>MCP (Model Context Protocol) is an open protocol proposed by Anthropic that lets AI Agents connect to external data sources and tools in a standardized way. Think of it as "the API for the AI era" — except MCP is self-describing (tools come with name, description, inputSchema), so AI Agents can automatically understand how to use them without integration docs.</p> <h2> See It in Action </h2> <p>Here's what a recruiter's AI Agent can do after connecting:</p> <p><strong>Query</strong>: Search for big data related experience<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight javascript"><code><span class="nf">search_resume</span><span class="p">({</span> <span class="na">query</span><span class="p">:</span> <span class="dl">"</span><span class="s2">big data</span><span class="dl">"</span> <span class="p">})</span> <span class="err">→</span> <span class="nx">Returns</span><span class="p">:</span> <span class="nx">relevant</span> <span class="nx">experience</span> <span class="nx">snippets</span> <span class="o">+</span> <span class="nx">source</span> <span class="nx">modules</span> <span class="o">+</span> <span class="nx">relevance</span> <span class="nx">scores</span> </code></pre> </div> <p><strong>Query</strong>: Is this candidate a fit for our role?<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight javascript"><code><span class="nf">evaluate_fit</span><span class="p">({</span> <span class="na">job_description</span><span class="p">:</span> <span class="dl">"</span><span class="se">\"</span><span class="s2">Senior full-stack engineer, 5+ years, React/Node.js,</span><span class="dl">"</span> <span class="nx">big</span> <span class="nx">data</span> <span class="nx">platform</span> <span class="nx">experience</span><span class="p">,</span> <span class="nx">DevOps</span><span class="dl">"</span><span class="s2"> }) → { score: 82, matched: [</span><span class="dl">"</span><span class="nx">react</span><span class="dl">"</span><span class="s2">, </span><span class="dl">"</span><span class="nx">node</span><span class="dl">"</span><span class="s2">, </span><span class="dl">"</span><span class="nx">big</span> <span class="nx">data</span><span class="dl">"</span><span class="s2">, </span><span class="dl">"</span><span class="nx">devops</span><span class="dl">"</span><span class="s2">, ...], missing: [</span><span class="dl">"</span><span class="nx">Kubernetes</span><span class="dl">"</span><span class="s2">], recommendation: </span><span class="dl">"</span><span class="nx">Strong</span> <span class="nx">match</span><span class="o">!</span> <span class="p">...</span><span class="dl">"</span><span class="s2"> } </span></code></pre> </div> <p><strong>Query</strong>: Give me a comprehensive assessment<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight javascript"><code><span class="nf">get_career_summary</span><span class="p">({})</span> <span class="err">→</span> <span class="p">{</span> <span class="nl">seniority</span><span class="p">:</span> <span class="dl">"</span><span class="s2">senior</span><span class="dl">"</span><span class="p">,</span> <span class="nx">totalYears</span><span class="p">:</span> <span class="mi">6</span><span class="p">,</span> <span class="nx">domains</span><span class="p">:</span> <span class="p">[</span><span class="dl">"</span><span class="s2">Big Data Cloud Services</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">Cross-platform Development</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">Indie Product Development</span><span class="dl">"</span><span class="p">],</span> <span class="nx">coreStrengths</span><span class="p">:</span> <span class="p">[</span><span class="dl">"</span><span class="s2">Multi-stack adaptability</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">Independent product delivery</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">AI protocol design</span><span class="dl">"</span><span class="p">],</span> <span class="nx">topSkills</span><span class="p">:</span> <span class="p">[{</span> <span class="na">name</span><span class="p">:</span> <span class="dl">"</span><span class="s2">React</span><span class="dl">"</span><span class="p">,</span> <span class="na">level</span><span class="p">:</span> <span class="mi">4</span><span class="p">,</span> <span class="na">yearsUsed</span><span class="p">:</span> <span class="mi">5</span> <span class="p">},</span> <span class="p">...]</span> <span class="p">}</span> </code></pre> </div> <p>All structured JSON — no PDF parsing, no web scraping. The AI Agent processes data programmatically.</p> <h2> The Key Insight Isn't Technical — It's Product Logic </h2> <h3> The Resume Submission Flow Has Changed </h3> <p>Old flow: Send PDF → HR skims manually → Maybe passes to AI for辅助判断</p> <p>New flow: Give HR a link + invite code → Their AI Agent connects via MCP → The Agent decides what to query and how to analyze</p> <p>The difference: <strong>the initiative shifts from HR to the AI Agent</strong>. The Agent queries dynamically based on what it cares about, rather than passively reading a fixed-format document.</p> <h3> Invite Code = Reverse Filtering </h3> <p>This is my favorite part of the design.</p> <p>When applying, you give HR a "link + invite code." HR needs to configure the invite code in their AI client before the Agent can access deep information.</p> <p>The process itself is a filter: <strong>teams without sufficient AI literacy can't even get through it</strong>. If they can, it means they're already using AI Agents in their hiring process — likely a more modern team.</p> <p>Plus, the invite code supports tracking — you can see exactly which tools the other party's Agent queried and when. Far more valuable than a "resume viewed" notification. You know <strong>what they actually care about</strong>.</p> <h3> Privacy Tiers </h3> <p>Public page and MCP responses expose different levels of information:</p> <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Tier</th> <th>What's visible</th> <th>Channel</th> </tr> </thead> <tbody> <tr> <td>Public</td> <td>Name, title, skill tags, work history</td> <td> <code>/</code> public page</td> </tr> <tr> <td>MCP Query</td> <td>Project details, skill proficiency, fit analysis</td> <td> <code>/mcp</code> (invite code required)</td> </tr> <tr> <td>Private</td> <td>Email, phone number</td> <td>Never exposed via MCP</td> </tr> </tbody> </table></div> <p>Not everything should be AI-accessible. Contact info belongs in the human communication phase.</p> <h2> Three-Layer Discovery Mechanism </h2> <p>How does an AI Agent know your resume has an MCP service? This is the most critical design decision — <strong>discovery matters more than connection</strong>.<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>AI Agent visits https://your-domain.com ↓ Parses HTML, discovers &lt;link rel="mcp" href="/mcp"&gt; ↓ Requests /.well-known/mcp for full metadata (tool list, auth method) ↓ Connects using invite code via Authorization header ↓ Starts querying </code></pre> </div> <p>Three layers of discovery, shallow to deep:</p> <p><strong>Layer 1: HTML Injection</strong> — Auto-injected into public page, zero maintenance cost<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight html"><code><span class="nt">&lt;link</span> <span class="na">rel=</span><span class="s">"mcp"</span> <span class="na">href=</span><span class="s">"/mcp"</span><span class="nt">&gt;</span> <span class="nt">&lt;link</span> <span class="na">rel=</span><span class="s">"llms-txt"</span> <span class="na">href=</span><span class="s">"/llms.txt"</span><span class="nt">&gt;</span> <span class="nt">&lt;script </span><span class="na">type=</span><span class="s">"application/ld+json"</span><span class="nt">&gt;</span> <span class="p">{</span> <span class="dl">"</span><span class="s2">@type</span><span class="dl">"</span><span class="p">:</span> <span class="dl">"</span><span class="s2">Person</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">name</span><span class="dl">"</span><span class="p">:</span> <span class="dl">"</span><span class="s2">...</span><span class="dl">"</span><span class="p">,</span> <span class="dl">"</span><span class="s2">knowsAbout</span><span class="dl">"</span><span class="p">:</span> <span class="p">[...]</span> <span class="p">}</span> <span class="nt">&lt;/script&gt;</span> </code></pre> </div> <p><strong>Layer 2: /.well-known/mcp</strong> — Standard well-known endpoint, machine-readable full MCP config<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight json"><code><span class="p">{</span><span class="w"> </span><span class="nl">"mcp"</span><span class="p">:</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="nl">"endpoint"</span><span class="p">:</span><span class="w"> </span><span class="s2">"/mcp"</span><span class="p">,</span><span class="w"> </span><span class="nl">"transport"</span><span class="p">:</span><span class="w"> </span><span class="s2">"http"</span><span class="p">,</span><span class="w"> </span><span class="nl">"auth"</span><span class="p">:</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="nl">"type"</span><span class="p">:</span><span class="w"> </span><span class="s2">"bearer"</span><span class="w"> </span><span class="p">},</span><span class="w"> </span><span class="nl">"tools"</span><span class="p">:</span><span class="w"> </span><span class="p">[</span><span class="err">...</span><span class="p">]</span><span class="w"> </span><span class="p">}</span><span class="w"> </span><span class="p">}</span><span class="w"> </span></code></pre> </div> <p><strong>Layer 3: /llms.txt</strong> — Follows the <a href="https://llmstxt.org/" rel="noopener noreferrer">llms.txt</a> standard, Markdown-formatted AI-friendly description</p> <p>All three layers are auto-generated server-side from the same <code>resume.json</code>. Update your resume, and all discovery info updates automatically.</p> <h2> Architecture </h2> <p>This is a self-hosted full-stack application, not SaaS. Single-user, deployed on your own server.<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>/ → Public page (visible to anyone, MCP discovery info injected) /mcp → MCP endpoint (invite code auth, for AI Agents) /edit → Resume editor (auth code protected, for yourself) </code></pre> </div> <p>Core principle: <strong>single source of truth</strong> — <code>resume.json</code>. Editor and MCP share the same data. MCP reads fresh data on every request — no cache staleness.</p> <p>The tech stack is nothing to brag about: React 19 + Vite 7 + Tailwind 4 for the editor, vanilla Node.js HTTP for the server, MCP SDK for the protocol layer, Zod for data validation. Search uses a hand-rolled TF scorer + mixed Chinese/English tokenization — good enough for keyword matching, nowhere near real semantic search.</p> <p>8 tools total:</p> <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Tool</th> <th>What it does</th> </tr> </thead> <tbody> <tr> <td><code>get_profile</code></td> <td>Public info, privacy fields never via MCP</td> </tr> <tr> <td><code>get_experience</code></td> <td>Work history, filterable by keyword</td> </tr> <tr> <td><code>get_projects</code></td> <td>Projects, filterable by tech stack</td> </tr> <tr> <td><code>get_skills</code></td> <td>Skill list with proficiency and years</td> </tr> <tr> <td><code>get_education</code></td> <td>Education background</td> </tr> <tr> <td><code>search_resume</code></td> <td>Full-text search, TF relevance scoring</td> </tr> <tr> <td><code>evaluate_fit</code></td> <td>Given a JD, outputs match score + missing skills</td> </tr> <tr> <td><code>get_career_summary</code></td> <td>Career overview: seniority, domains, core strengths</td> </tr> </tbody> </table></div> <p>Each tool supports <code>format: 'text' | 'json'</code> — text for humans, json for programs.</p> <h2> Engineering Decisions Worth Mentioning </h2> <p><strong>Atomic writes instead of a database</strong> — Resume data is ~2KB. Even SQLite feels overkill. write-to-temp + rename is atomic, plus a write queue for concurrency safety.</p> <p><strong>Fresh MCP Server per request</strong> — Instead of a long-lived MCP Server created at startup, each request calls <code>createMcpServer(currentResume)</code>. This guarantees the latest data — edits are immediately visible to MCP.</p> <p><strong>Zod schema validation</strong> — Validate on both read and write. The frontend editor validates too, but never trusting client input is basic hygiene.</p> <p><strong>stdio mode</strong> — Besides HTTP, supports <code>start:stdio</code> for direct Claude Desktop configuration. No server needed for local development.</p> <h2> Deployment </h2> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code><span class="c"># One-command Docker deployment</span> <span class="nb">cp </span>config.example.json config.json <span class="c"># Set your authCode</span> docker compose up <span class="nt">-d</span> </code></pre> </div> <p>Or connect directly via Claude Desktop:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight json"><code><span class="p">{</span><span class="w"> </span><span class="nl">"mcpServers"</span><span class="p">:</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="nl">"resume"</span><span class="p">:</span><span class="w"> </span><span class="p">{</span><span class="w"> </span><span class="nl">"command"</span><span class="p">:</span><span class="w"> </span><span class="s2">"npx"</span><span class="p">,</span><span class="w"> </span><span class="nl">"args"</span><span class="p">:</span><span class="w"> </span><span class="p">[</span><span class="s2">"tsx"</span><span class="p">,</span><span class="w"> </span><span class="s2">"server/stdio.ts"</span><span class="p">],</span><span class="w"> </span><span class="nl">"cwd"</span><span class="p">:</span><span class="w"> </span><span class="s2">"/path/to/resume-maker"</span><span class="w"> </span><span class="p">}</span><span class="w"> </span><span class="p">}</span><span class="w"> </span><span class="p">}</span><span class="w"> </span></code></pre> </div> <h2> Honest Take </h2> <p>The search and matching algorithms in this project are honestly mediocre. TF scoring is just keyword frequency counting, skill normalization is a lookup table, and evaluate_fit's matching logic is essentially string containment checking. These are far from "precisely understanding personal data."</p> <p>But the interesting part isn't the implementation.</p> <p>What's interesting is the new possibilities created by the <strong>Resume + MCP</strong> combination:</p> <ol> <li> <strong>Information access has changed</strong> — From "reading a fixed document" to "AI querying structured data on demand"</li> <li> <strong>Filtering logic has changed</strong> — Invite codes reverse-filter recruiters by their AI literacy</li> <li> <strong>Data ownership has changed</strong> — Data lives on your own server, you control who sees what</li> </ol> <p>These three shifts have nothing to do with implementation complexity. They're innovations at the <strong>product form</strong> level.</p> <p>MCP is still early — almost no recruiters are actually using AI Agents to connect to MCP endpoints. But the direction is right. As AI becomes more involved in hiring, resumes should evolve from "documents for humans" into "data sources for AI."</p> <p>If you're also experimenting with MCP, or interested in "AI-native personal information presentation," let's connect.</p> <p><strong>Live Demo</strong>: <a href="http://47.97.101.194:965" rel="noopener noreferrer">http://47.97.101.194:965</a> (Invite code: <code>qm8vvf</code> — connect to the MCP endpoint to try all tools)</p> <p><strong>Project</strong>: <a href="https://github.com/Leeson-Wong/resume-maker" rel="noopener noreferrer">github.com/Leeson-Wong/resume-maker</a></p> <p>If you find it interesting, a Star would be appreciated.</p> <h2> Another Project: BrickLife </h2> <p>Quick shoutout to another thing I built — a simulation/management mini-game called "BrickLife"</p> <p>Play: <strong><a href="http://47.97.101.194" rel="noopener noreferrer">http://47.97.101.194</a></strong> (Best experience on mobile)</p> <blockquote> <p>This article was co-written with Claude Code + GLM-5.1.</p> </blockquote>