Community

Join our groups for discussion and support:

Recent Updates

v0.5.0 · Desktop Client

• Desktop client released: macOS (Apple Silicon / Intel) and Windows x64 installers — zero config, no account, data stays local
• Built-in update notice that downloads the installer for your machine
• Web and desktop now ship together: one tag produces both the Docker image and all three installers
• Fixed PDF export failing after the container had been up for a while (#95)
• Resume items can be moved up/down and inserted above (#89, thanks @Silas-Zhu)

v0.3.4 · Brand Color System & Theme Switching

• Introduced semantic --brand-* CSS tokens; replaced hardcoded pink-* across 60+ files
• New brand switcher in the user menu with three presets: Mint (default), Blue, Pink
• SSR-safe anti-flicker hydration; legacy values auto-migrated via localStorage
• Added a Mint resume preset to the theme editor
• Export pipelines (PDF / HTML / DOCX) now read from src/lib/brand-constants.ts

v0.3.3 · Mobile Experience & Interview Report Stability

• Mobile-friendly template preview and bottom action bars for preview/share pages
• Fixed mobile scroll: canvas/preview roots use h-full instead of flex-1
• Improved stability of interview report generation

v0.3.2 · Runtime Environment Variables

• Removed all NEXT_PUBLIC_* build-time variables in favor of runtime env

v0.3.1 · Auth Runtime Flag

• Changed NEXT_PUBLIC_AUTH_ENABLED to runtime AUTH_ENABLED

Screenshots

Template Gallery	Resume Editor

AI Resume Generation	AI Resume Parsing (Image)

AI Optimization	Grammar Check

Grammar Auto-Fix	JD Match Analysis

Multi-Format Export	Share Link

Shared Resume Page	AI Professional Photo

QR Code Section

Interview Setup	Mock Interview

Interview List	Interview Report

Deployment Video

Watch the full deployment walkthrough on Bilibili:

Watch on Bilibili →

Features

Resume Editing

• Drag & Drop Editor — Visually arrange and reorder resume sections and items
• Inline Editing — Click any field to edit directly on the canvas
• 50 Professional Templates — Classic, Modern, Minimal, Creative, ATS-Friendly, Timeline, Nordic, Swiss, and more
• Theme Customization — Colors, fonts, spacing, and margins with live preview
• Undo / Redo — Full edit history (up to 50 steps)
• Auto Save — Configurable interval (0.3s–5s), with manual save option
• Markdown Support — Use Markdown syntax in text fields to format content (e.g., **bold** for bold text)

Markdown Formatting

The following resume sections support Markdown syntax:

Supported syntax:

**bold text**    → bold
`code text`      → inline code
- item           → bullet list

Skills, Certifications, and Personal Info fields do not support Markdown.

AI Capabilities

• AI Chat Assistant — Conversational AI integrated in the editor, with multi-session support and persistent history
• AI Resume Generation — Generate a complete resume from job title, experience, and skills
• Resume Parsing — Upload an existing PDF or image, AI extracts all content automatically
• JD Match Analysis — Compare resume against a job description: keyword matching, ATS score, and improvement suggestions
• Cover Letter Generation — AI-tailored cover letter based on resume and JD, with tone selection (formal / friendly / confident)
• Grammar & Writing Check — Detect weak verbs, vague descriptions, and grammar issues; returns a quality score
• Translation — Translate resume content across 10 languages while preserving technical terms
• Flexible AI Provider — Supports OpenAI, Anthropic, and any OpenAI-compatible endpoint (such as OrcaRouter, which exposes 200+ models behind one key); each user configures their own key in-app

Mock Interview

• JD-Based Interview Simulation — Paste a job description, AI plays different interviewer roles in sequence
• 6 Preset Interviewers — HR, Technical, Scenario, Behavioral, Project Deep Dive, Leader — each with unique personality and questioning style
• Custom Interviewers — Create your own interviewer with custom focus areas and style
• Smart Follow-ups — AI adapts questions based on answer quality, probing deeper when needed
• Interview Controls — Skip questions, request hints, mark for review, pause/resume
• Detailed Report — Per-question scoring, competency radar chart, improvement plan with resources
• History Comparison — Track score trends and dimension progress across interviews
• PDF & Markdown Export — Export interview reports for offline review

Export & Sharing

• Multi-Format Export — PDF (Puppeteer + Chromium), Smart One-Page PDF (auto-fit to single page), DOCX, HTML, TXT, JSON
• JSON Import — Import a previously exported JSON file to restore or create a resume; supported both in the editor (overwrite current) and on the dashboard (create new)
• Link Sharing — Token-based shareable links with optional password protection
• View Counter — Track how many times a shared resume has been viewed

Management

• Multi-Resume Dashboard — Grid and list views, search, sort (by date, name)
• Import from JSON — Create a new resume from a JSON export directly on the dashboard
• Duplicate & Rename — Quick resume management actions
• Interactive Tours — Step-by-step onboarding for first-time users

Other

• Bilingual UI — Full Chinese (zh) and English (en) interface
• Dark Mode — Light, dark, and system theme support
• Flexible Auth — Google OAuth or browser fingerprint (zero-config)
• Local SQLite Storage — Zero-config, file-based database

Tech Stack

Desktop Client

Try the desktop app — nothing to deploy, no account, no network. Download, install, start writing.

It runs the same application as the web version — all 50 templates, AI polish, JD matching, mock interviews, every export format. The difference is only where it runs:

• Zero config — no Docker, no database, no AUTH_SECRET. Install and open.
• No account — one local user, no sign-in, no fingerprinting
• Your data stays put — resumes live in a SQLite file on your machine and never reach a server
• Your own AI key — set it in Settings → AI; requests go straight to the provider you configured
• Update notices — a small panel appears when a new version is out and downloads the installer built for your machine

The only outbound request it makes is asking GitHub whether a newer version exists, and that can be turned off (see Updates below).

Download

Grab the build for your system from the latest release:

First launch

The installers carry an ad-hoc signature — valid, but anonymous. There is no Apple Developer ID and no notarization, so the OS blocks the first launch. This is expected.

macOS. Run this once, then open the app normally:

xattr -dr com.apple.quarantine /Applications/JadeAI.app

Or via the GUI: after the launch is blocked, open System Settings → Privacy & Security and click Open Anyway near the bottom.

Do not rely on right-click → Open. macOS 15 removed that bypass for un-notarized apps; it no longer does anything.

If you see "JadeAI is damaged and can't be opened" rather than "cannot be verified", the download is incomplete or was tampered with — re-download it. A correctly built release reports "cannot be verified", which is clearable.

Windows. On the SmartScreen prompt, click More info → Run anyway.

Where your data lives

jade.db is the SQLite database; jade-settings.json holds window state and preferences. Uninstalling the app does not delete either — remove the folder to erase your data.

Updates

On launch the app asks GitHub once whether a newer version exists. If there is one, a small panel appears in the corner of the window — not a modal dialog — offering to download the installer built for your machine, so you don't pick between three files. Progress shows in the panel; when it finishes you can open the installer or reveal it in your file manager. It can be collapsed, dismissed, or silenced for that version.

It stops there: installing is still a manual step (drag to Applications, or run the .exe). Silent install would need Squirrel, which refuses an ad-hoc signed app. Your data is untouched by an update.

This is the only outbound request the app makes. To turn it off, set "updateCheckEnabled": false in jade-settings.json (see the paths above) and restart.

Getting Started

Docker (Recommended)

# Generate a secret key first
openssl rand -base64 32

docker run -d -p 3000:3000 \
  -e AUTH_SECRET=<your-generated-secret> \
  -v jadeai-data:/app/data \
  twwch/jadeai:latest

Open http://localhost:3000. Database auto-migrates and seeds on first start.

AUTH_SECRET is required for session encryption. Generate one with openssl rand -base64 32.

AI Configuration: No server-side AI env vars needed. Each user configures their own API Key, Base URL, and Model in Settings > AI within the app. If you don't have a provider key yet, OrcaRouter gives you 200+ models through a single OpenAI-compatible endpoint — see AI provider setup.

docker run -d -p 3000:3000 \
  -e AUTH_ENABLED=true \
  -e AUTH_SECRET=your-secret \
  -e GOOGLE_CLIENT_ID=xxx \
  -e GOOGLE_CLIENT_SECRET=xxx \
  -v jadeai-data:/app/data \
  twwch/jadeai:latest

Local Development

Prerequisites

• Node.js 18+
• pnpm 9+

Installation

git clone https://github.com/twwch/JadeAI.git
cd JadeAI

pnpm install
cp .env.example .env.local

Configure Environment

Edit .env.local:

# Auth (defaults to fingerprint mode, no config needed)
AUTH_ENABLED=false

AI Configuration: No server-side env vars needed. Each user configures their own API Key, Base URL, and Model in Settings > AI within the app. See AI provider setup for using OrcaRouter to access 200+ models with one key.

See .env.example for all available options (Google OAuth, custom SQLite path, etc.).

Initialize Database & Run

# Generate and run migrations
pnpm db:generate
pnpm db:migrate

# (Optional) Seed with sample data
pnpm db:seed

# Start dev server
pnpm dev

Open http://localhost:3000.

AI Provider Setup

All AI features (chat, resume generation, JD matching, mock interview, translation…) need an API key. JadeAI never stores keys on the server — you fill them in under Settings > AI in the app, and they stay in your browser's local storage.

Any OpenAI-compatible endpoint works. Point it at OpenAI, Anthropic, a self-hosted gateway, or a router.

Using OrcaRouter (200+ models, one key)

OrcaRouter is an OpenAI-compatible AI gateway: one key gives you access to 200+ models (OpenAI, Anthropic, Gemini, DeepSeek, Qwen and more) at provider pricing with no per-token markup, plus automatic failover between providers. Handy if you'd rather not sign up with every vendor separately, or want to switch models inside JadeAI without swapping keys.

1. Register at orcarouter.ai (free tier, no credit card) and create an API key.
2. In JadeAI, open Settings > AI and fill in:

3. Save, then use any AI feature in the editor.

To use OpenAI or Anthropic directly instead, just fill in that provider's base URL, key, and model ID in the same panel.

Environment Variables

Scripts

Project Structure

src/
├── app/                        # Next.js App Router
│   ├── [locale]/               # i18n routes (/zh/..., /en/...)
│   │   ├── dashboard/          # Resume list & management
│   │   ├── editor/[id]/        # Resume editor
│   │   ├── preview/[id]/       # Full-screen preview
│   │   ├── templates/          # Template gallery
│   │   └── share/[token]/      # Public shared resume viewer
│   └── api/
│       ├── ai/                 # AI endpoints
│       │   ├── chat/           #   Streaming chat with tool calls
│       │   ├── generate-resume/#   AI resume generation
│       │   ├── jd-analysis/    #   JD match analysis
│       │   ├── grammar-check/  #   Grammar & writing check
│       │   ├── cover-letter/   #   Cover letter generation
│       │   ├── translate/      #   Resume translation
│       │   └── models/         #   List available AI models
│       ├── resume/             # Resume CRUD, export, parse, share
│       ├── share/              # Public share access
│       ├── user/               # User profile & settings
│       └── auth/               # NextAuth handlers
├── components/
│   ├── ui/                     # shadcn/ui base components
│   ├── editor/                 # Editor canvas, sections, fields, dialogs
│   ├── ai/                     # AI chat panel & bubble
│   ├── preview/templates/      # 50 resume templates
│   ├── dashboard/              # Dashboard cards, grid, dialogs
│   └── layout/                 # Header, theme provider, locale switcher
├── lib/
│   ├── db/                     # Schema, repositories, migrations, adapters
│   ├── auth/                   # Auth configuration
│   └── ai/                     # AI prompts, tools, model config
├── hooks/                      # Custom React hooks (7 hooks)
├── stores/                     # Zustand stores (resume, editor, settings, UI, tour)
└── types/                      # TypeScript type definitions

Templates

JadeAI includes 50 professionally designed resume templates covering a wide range of styles and industries:

API Reference

Contributing

Contributions are welcome! Here's how to get started:

1. Fork the repository
2. Create a feature branch: git checkout -b feat/your-feature
3. Commit your changes: git commit -m 'feat: add your feature'
4. Push to the branch: git push origin feat/your-feature
5. Open a Pull Request

FAQ

Star History

License

Apache License 2.0