ResumeAgent

KO | EN

gitlite — search

#python #hacktoberfest #typescript #llm #golang #nodejs #windows #awesome #cli #kubernetes #pytorch #security #ios #mcp

(★ 4)

ResumeAgent are Purpose-built AI agents for the modern job search. Tailor to any role, transform to a polished PDF, and analyze your fit.

ResumeAgent 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

<p align="center">
  <h1 align="center">ResumeAgent</h1>
  <h3 align="center">
    An open-source, agentic resume toolkit.
  </h3>
</p>

<h4 align="center">
  Grab a coffee ☕️ and let the agents handle your resume.
</h4>

<p align="center">
  <a href="#quick-start">Quick Start</a> •
  <a href="#docker-deployment">Docker Deploy</a> •
  <a href="#features">Features</a> •
  <a href="#architecture">Architecture</a> •
  <a href="#docker-latex-sandbox">Docker</a> •
  <a href="#updating">Updating</a>
</p>

---

## Why ResumeAgent?

ResumeAgent is an **open-source** project that leverages a coding agent plus a library of composable **Skills** to make resume workflows **agentic** and dramatically more efficient. Instead of a black-box “rewrite my resume”, it decomposes parsing, tailoring, scoring, and heatmapping into transparent, inspectable steps — so you can customize prompts/strategies, plug in new skills, and automate end-to-end resume work with confidence.

**BYOK (Bring Your Own Key)** — no vendor lock-in, no subscriptions. Plug in your LLM API key and own your entire workflow.

---

## Features

### 📄 Resume Transformer

Convert any PDF/DOCX resume into a beautifully formatted LaTeX PDF. AI-powered parsing extracts structured data, then fills your chosen template with intelligent section mapping.

<details open>
<summary>🎬 Watch Demo</summary>
<br>
<video src="https://github.com/user-attachments/assets/aa2f7435-f36d-410b-9719-8a2d71624750" controls width="100%"></video>
</details>

- Upload PDF, DOCX, or images (OCR via Tesseract.js)
- 4 LaTeX templates: **Basic**, **Modern**, **Classic**, **Academic**
- Live TeX editor with instant PDF re-render
- Full CJK / multilingual support

### ✂️ Resume Tailor

Optimize your resume for a specific job description through block-by-block AI improvements. Zero-fabrication policy — only reframes your existing experience.

<details open>
<summary>🎬 Watch Demo</summary>
<br>
<video src="https://github.com/user-attachments/assets/a904bd0b-031d-4c7f-9b7b-d97c596c3378" controls width="100%"></video>
</details>

- JD analysis → match scoring → targeted optimization
- Strategies: keyword alignment, gap bridging, quantification, reordering, pruning
- Real-time streaming progress with per-block diffs
- LaTeX-rendered PDF + DOCX output

### 🔥 Resume Heatmap

Visually analyze how well a resume aligns with a job description. Character-level colorization shows exactly what matters.

<details open>
<summary>🎬 Watch Demo</summary>
<br>
<video src="https://github.com/user-attachments/assets/9d7c99fc-4e85-4530-831d-f3671729eccc" controls width="100%"></video>
</details>

- Three-level analysis: block → element → phrase
- Color-coded relevance: 🔴 high match · 🟠 partial · 🟢 irrelevant
- Employer-facing candidate evaluation tool
- Rendered as a colorized LaTeX PDF overlay

### 🔍 Resume Parser

Extract structured data from any resume using custom schemas. Define exactly what you want to extract with a visual builder or raw JSON editor.

<details open>
<summary>🎬 Watch Demo</summary>
<br>
<video src="https://github.com/user-attachments/assets/e394f1ce-2616-4e55-ac68-1bddbd2016ff" controls width="100%"></video>
</details>

### 🎨 Template Gallery

Browse and preview all available resume templates with live PDF previews and one-click download.

---

## Quick Start

### One-command setup (recommended for development)

The setup script checks your environment (Node.js, pnpm, Docker), installs missing tools, pulls the LaTeX Docker image, installs dependencies, and creates the `.env` file — all in one go.

```bash
git clone https://github.com/ApplyU-ai/ResumeAgent.git
cd ResumeAgent
./setup.sh
```

Then choose a mode:

```bash
# Development — hot-reload, instant feedback
pnpm dev

# Production — compile once, then serve
pnpm build
pnpm start
```

Open [http://localhost:3000](http://localhost:3000) in your browser.

### Docker Deployment

Deploy ResumeAgent as a single Docker container

```bash
# Pull pre-built image and run (fastest)
docker run -p 3000:3000 ghcr.io/applyu-ai/resumeagent/resumeagent:latest

# Or build locally with docker compose
docker compose up -d

# Or build and run manually
docker build -t resumeagent .
docker run -p 3000:3000 resumeagent
```

Open [http://localhost:3000](http://localhost:3000).

### Manual setup

<details>
<summary>Click to expand manual steps</summary>

#### Prerequisites

- **Node.js** >= 20
- **pnpm** >= 10 (`npm install -g pnpm`)
- **Docker** — installed and running

#### Steps

```bash
# 1. Clone & install
git clone https://github.com/ApplyU-ai/ResumeAgent.git
cd ResumeAgent
pnpm install

# 2. Pull LaTeX Docker image
docker pull ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest
docker tag ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest resumeagent-latex

# 3. Configure environment
cp packages/backend/.env.example packages/backend/.env

# 4a. Development (hot-reload)
pnpm dev

# 4b. Production (build then serve)
pnpm build
pnpm start
```

</details>

### Environment variables

| Variable | Default | Description |
|----------|---------|-------------|
| `PORT` | `3001` | Backend server port |
| `CORS_ORIGINS` | `http://localhost:3000` | Comma-separated allowed origins |
| `FILE_STORE_PATH` | `./tmp/files` | Uploaded file storage path |
| `SANDBOX_MODE` | `auto` | `auto` (detect), `docker` (sandbox), or `direct` (host TeX Live) |

> **Note:** LLM API keys are provided per-request via the UI — no server-side API key configuration needed.

### Supported LLM providers & models

- **Anthropic**: `claude-opus-4-5`, `claude-opus-4-6`, `claude-sonnet-4-6`
- **OpenAI**: `gpt-5.2`, `gpt-5.1`, `gpt-5`
- **Azure OpenAI**: `gpt-5.1`, `gpt-5.2` 
- **Google**: `gemini-3-flash-preview`, `gemini-3-pro-preview`, `gemini-3.1-pro-preview`

---

## Tech Stack

| Layer | Technology |
|-------|-----------|
| Frontend | Next.js 16, React 19, Tailwind CSS v4, Vercel AI SDK |
| Backend | Hono, Node.js, [pi-mono](https://github.com/badlogic/pi-mono) agent framework |
| Shared | TypeBox runtime schemas, shared between frontend & backend |
| LaTeX Sandbox | Docker (TeX Live + XeLaTeX), sandboxed with security constraints |
| Monorepo | pnpm workspaces + Turborepo |
| Linting | Biome |
| Testing | Vitest |
| CI/CD | GitHub Actions |

---

## Architecture

```
┌──────────────────────────────────────────────────────────────────────┐
│                        Next.js Frontend                              │
│                                                                      │
│  ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ ┌────────────┐  │
│  │ Transformer  │ │ Tailor       │ │ Heatmap      │ │ Parser     │  │
│  │ /transformer │ │ /resume-     │ │ /resume-     │ │ /resume-   │  │
│  │              │ │  tailor      │ │  heatmap     │ │  parser    │  │
│  └──────┬───────┘ └──────┬───────┘ └──────┬───────┘ └─────┬──────┘  │
│         └────────────────┼────────────────┼────────────────┘         │
│                          │ SSE / REST                                │
└──────────────────────────┼───────────────────────────────────────────┘
                           │
┌──────────────────────────┼───────────────────────────────────────────┐
│                     Hono Backend (port 3001)                         │
│                                                                      │
│  ┌───────────────────────────────────────────────────────────────┐   │
│  │                    Agent Orchestrator                          │   │
│  │                 (pi-mono / pi-agent-core)                      │   │
│  │                                                                │   │
│  │  ┌─────────────┐ ┌─────────────┐ ┌─────────────────────────┐  │   │
│  │  │ resume-parse│ │ jd-analyze  │ │ template-fill           │  │   │
│  │  ├─────────────┤ ├─────────────┤ ├─────────────────────────┤  │   │
│  │  │ match-score │ │ block-      │ │ identify-optimization-  │  │   │
│  │  │             │ │ optimize    │ │ targets                 │  │   │
│  │  ├─────────────┤ ├─────────────┤ ├─────────────────────────┤  │   │
│  │  │ heatmap-    │ │ tex-heatmap-│ │ latex-compile           │  │   │
│  │  │ analyze     │ │ colorize    │ │ (Docker sandbox)        │  │   │
│  │  └─────────────┘ └─────────────┘ └─────────────────────────┘  │   │
│  └───────────────────────────────────────────────────────────────┘   │
│                                │                                     │
│                    ┌───────────┴───────────┐                         │
│                    │    Docker Sandbox      │                         │
│                    │  (resumeagent-latex)   │                         │
│                    └───────────────────────┘                         │
└──────────────────────────────────────────────────────────────────────┘
```

### Agent Skills

ResumeAgent uses a modular **Skill** architecture. Each skill is a self-contained unit with its own prompt, I/O schema, and optional scripts:

| Skill | Type | Description |
|-------|------|-------------|
| `resume-parse` | LLM | Extract structured ResumeData from PDF/DOCX |
| `jd-analyze` | LLM | Analyze job descriptions into structured requirements |
| `match-score` | LLM | Score resume-JD alignment (0–100) with per-category breakdown |
| `identify-optimization-targets` | LLM | Prioritize which resume blocks to optimize |
| `block-optimize` | LLM | Optimize a single block (keyword, quantification, reorder, prune) |
| `template-fill` | LLM | Fill LaTeX templates with parsed resume data |
| `heatmap-analyze` | LLM | Per-block relevance analysis at phrase level |
| `tex-heatmap-colorize` | Deterministic | Inject xcolor highlights into LaTeX source |
| `latex-compile` | Deterministic | Compile TeX → PDF (+ optional DOCX) via Docker sandbox |
| `schema-parse` | LLM | Custom-schema extraction from resumes |

---

## Docker LaTeX Sandbox

ResumeAgent compiles LaTeX inside a **sandboxed Docker container** with strict security constraints:

### Pull pre-built image (recommended)

```bash
docker pull ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest
docker tag ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest resumeagent-latex
```

### Build locally

```bash
cd sandbox
docker build -t resumeagent-latex .
```

The image includes TeX Live (XeLaTeX + pdflatex), CJK fonts (Noto CJK), and pdf2docx for optional DOCX generation.

### Sandbox security

The Docker sandbox runs with strict isolation:

- `--network=none` — no network access
- `--read-only` — read-only root filesystem
- `--memory=256m` — 256 MB memory cap
- `--cpus=1` — single CPU
- `--pids-limit=50` — process limit
- Non-root user (`latex`)

---

## Updating

When a new version of ResumeAgent is released:

```bash
# 1. Pull latest code
git pull origin main

# 2. Re-run setup (installs new deps, updates Docker image if needed)
./setup.sh

# 3a. Development
pnpm dev

# 3b. Production
pnpm build
pnpm start
```

To update **only** the LaTeX Docker image:

```bash
docker pull ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest
docker tag ghcr.io/applyu-ai/resumeagent/resumeagent-latex:latest resumeagent-latex
```

---

## Individual Package Commands

```bash
# ── Development (hot-reload) ──────────────────────────────
pnpm dev                                  # all packages
pnpm --filter @resumeagent/frontend dev   # frontend only
pnpm --filter @resumeagent/backend dev    # backend only

# ── Production ───────────────────────────────────────────
pnpm build    # compile all packages
pnpm start    # serve built output

# ── Quality ──────────────────────────────────────────────
pnpm check                          # TypeScript type-check
pnpm test                           # Vitest unit tests
npx @biomejs/biome check --write .  # lint & format (Biome)
```

---

## Acknowledgements

- [pi-mono](https://github.com/badlogic/pi-mono)
- [Vercel AI SDK](https://github.com/vercel/ai)
- [Hono](https://hono.dev)
- [TeX Live](https://www.tug.org/texlive/)

---

## License

[Apache-2.0](./LICENSE)

---

## Connect

For more information or questions, please contact zijian.ling@applyu.ai

ResumeAgent

ResumeAgent 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

ResumeAgent

ResumeAgent 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README