careervivid · CLI

Your AI-powered career terminal — autonomous job applications, resume editing, job pipeline tracking, and portfolio publishing from the command line.

Table of Contents

• Quick Start
• Commands
  • cv agent ⭐ AI Agent
  • cv agent --jobs 🤖 Autonomous Job Applications
  • cv agent --resume 📄 Resume CRUD
  • cv jobs
  • cv publish
  • cv whiteboard
  • cv profile
  • cv auth
  • cv login
  • cv config
• AI Credits
• Bring Your Own API Key (BYO)
• browser-use Setup (Autonomous Apply)
• Whiteboard Templates
• AI Agent Integration (MCP / CI)
• Updating
• Troubleshooting

Quick Start

# 1. Install globally
npm install -g careervivid

# 2. Log in and get your free API key
cv login
# → opens careervivid.app/developer in your browser
# → copy your API key, then:
cv auth set-key cv_live_YOUR_KEY_HERE

# 3. Start the AI agent
cv agent

# 4. Or jump straight into job-hunting mode
cv agent --jobs

Free tier includes 100 AI credits/month — no credit card required.

Commands

cv agent

An autonomous AI agent that runs interactively in your terminal. Choose from CareerVivid Cloud (credits deducted from your account) or Bring Your Own API Key.

cv agent                  # General-purpose agent (Gemini Flash Lite)
cv agent --resume         # Resume CRUD mode: read, update fields, tailor
cv agent --jobs           # Job-hunting mode: search, apply, track pipeline
cv agent --coding         # Full coding toolkit: file I/O, shell execution, search
cv agent --pro            # Gemini Pro with extended thinking

Options:

Configure your default provider:

cv agent config
# → Interactive wizard: pick provider, model, API key

cv agent --jobs

The most powerful mode. A conversational agent that thinks like a senior recruiter, manages your pipeline, and can autonomously fill out job applications using browser-use.

cv agent --jobs

What it can do in a conversation:

Autonomous Apply Harness (apply_to_job tool):

The agent uses a multi-stage harness to fill job applications:

1. URL Verification — Every job link is verified live before the agent acts. Broken, hallucinated, or redirect-to-homepage URLs are rejected automatically.
2. ATS Detection — Automatically detects Ashby, Greenhouse, Lever, Workday, iCIMS, and more.
3. TypeScript Adapter — Fast-fills known fields (firstName, email, phone, LinkedIn, GitHub) using pre-loaded profile data.
4. browser-use AI handoff — For complex fields (cover letter, work auth, custom questions), the browser-use agent takes over with a rich task prompt built from your resume.
5. Manual review gate — The browser stays open. The agent never clicks Submit — you review and submit yourself.

Requires browser-use setup — see browser-use Setup.

Pipeline tracking (local CSV):

The jobs agent also manages a local jobs.csv pipeline (v2 schema) with 8 attention metrics:

Priority Score = 40% attention + 30% excitement + 20% fit + 10% recency

cv agent --resume

Full CRUD control over your CareerVivid resume from the terminal.

cv agent --resume

Available tools:

Supported fields for set_resume_fields:

firstName, lastName, email, phone
address, city, state, country, postalCode
jobTitle, currentCompany
linkedin, github, portfolio
summary (professional summary)
addSkill / removeSkill
resumeTitle (document name)

Example conversation:

You:   Set my last name to Zhu and my phone to +1-408-599-4164
Agent: ✅ Resume updated — review at https://careervivid.app/edit/...

You:   Add Python and TypeScript to my skills
Agent: ✅ Skills updated

You:   Tailor my resume for this job description: [paste JD]
Agent: ✅ Tailored copy created — https://careervivid.app/edit/...

cv jobs

Non-interactive job commands (use cv agent --jobs for a conversational experience).

cv jobs hunt --role "Software Engineer" --score 60
cv jobs apply https://jobs.ashbyhq.com/openai/...
cv jobs list
cv jobs update

cv publish

Publish a Markdown article or Mermaid diagram to your CareerVivid portfolio.

cv publish <file> [options]
cv publish -       # read from stdin

Privacy default: All articles published via cv publish or cv agent are private by default. Use --public to make them visible to the community.

cv publish article.md --title "How I Built a CLI" --tags "node,typescript"
cv publish article.md --public               # share with community
cv publish architecture.mmd --title "System Architecture"
cat writeup.md | cv publish - --title "Architecture Breakdown" --json

cv whiteboard

Create and publish Mermaid architecture diagrams.

cv new my-diagram --template system-arch
cv list-templates

cv profile

Manage your CareerVivid profile.

cv auth

Manage your CareerVivid API key. Get your key at careervivid.app/developer.

cv auth set-key cv_live_YOUR_KEY_HERE
cv auth check
# ✔  Authenticated as Jiawen Zhu (jiawen@careervivid.app)

# Or use env var without saving locally
CV_API_KEY=cv_live_YOUR_KEY_HERE cv publish article.md

cv login

Open the CareerVivid sign-in page and interactively save your API key.

cv login

cv config

View and modify CLI configuration stored at ~/.careervividrc.json.

Available keys:

cv config set targetCompanies "OpenAI, Google, Vercel"
cv config set llmProvider openai
cv config set llmModel gpt-4o

AI Credits

Credit costs per agent turn:

Bring Your Own Key: Using --provider openai (or any non-CareerVivid provider) deducts zero credits — you pay your provider directly.

Get your key and check balance at careervivid.app/developer.

Bring Your Own API Key (BYO)

cv agent --provider openai --model gpt-4o --api-key sk-...
cv agent --provider anthropic --model claude-opus-4-5 --api-key sk-ant-...
cv agent --provider openrouter --model mistralai/mistral-7b-instruct --api-key sk-or-...

# Save a provider permanently
cv agent config

browser-use Setup (Autonomous Apply)

The apply_to_job tool in cv agent --jobs and cv jobs apply use browser-use — a Python AI browser automation library — to fill job applications.

Prerequisites

• Python 3.11+ with pip
• A virtual environment (recommended)

Setup (one-time)

# 1. Create a Python virtual environment
mkdir -p ~/careervivid/browser-use
cd ~/careervivid/browser-use
python3 -m venv .venv
source .venv/bin/activate

# 2. Install browser-use
pip install browser-use

# 3. Install Playwright browsers
playwright install chromium

# 4. Verify
python -c "import browser_use; print('browser-use ready')"

How it works

When you run cv agent --jobs and ask to apply to a job:

1. The agent calls apply_to_job with the job URL
2. The CLI spawns src/apply/browser_sidecar.py (bundled with the npm package)
3. The sidecar launches a real Chrome browser (non-headless) using your profile
4. An AI agent fills the form field-by-field using your resume data
5. The browser stays open — you review every field and submit manually

The agent will never click the final Submit button. You remain in full control.

Resume PDF

Place your resume at ~/.careervivid/resume.pdf and the sidecar will upload it automatically to resume upload fields.

cp ~/Downloads/MyResume.pdf ~/.careervivid/resume.pdf

Whiteboard Templates

Run cv list-templates to see all available templates:

AI Agent Integration (MCP / CI)

Use --json for structured output in automated pipelines.

# Pipe AI-generated markdown directly
echo "# My Architecture\n\nExplains the new service..." \
  | cv publish - --title "New Service Explained" --tags "architecture" --json

# Output:
# [{ "postId": "abc123", "url": "https://careervivid.app/community/post/abc123" }]

System prompt for Cursor / Claude / Gemini:

You have access to the `cv` CLI tool.

Available commands:
- cv publish <file>               Publish an article (private by default, use --public to share)
- cv agent --resume               Resume CRUD: read, update fields, tailor
- cv agent --jobs                 Job hunting + autonomous apply
- cv jobs hunt --role "..."       AI job search
- cv jobs apply <url>             Autonomous form filling (stops before submit)
- cv jobs list                    View job tracker

Rules:
1. Never include real API keys in published content.
2. Always use --json for machine-readable output.
3. Articles are PRIVATE by default. Only use --public if user explicitly requests it.

Updating

npm install -g careervivid
cv -v

Troubleshooting

cv: command not found

npm config get prefix
# Add <prefix>/bin to your PATH in ~/.zshrc or ~/.bashrc

Unauthorized error

cv auth check
CV_API_KEY=cv_live_YOUR_KEY cv publish article.md

browser_sidecar.py not found

# browser-use is not set up. Follow the browser-use Setup section above.
# Verify your venv path:
ls ~/careervivid/browser-use/.venv/bin/python

aiohttp is not defined or ainvoke() takes 2 positional arguments

# Update to the latest version (fixed in v1.12.5+)
npm install -g careervivid@latest

❌ No API key found when using CareerVivid Cloud

cv auth check          # verify your CV API key is set
cv agent config        # reconfigure your LLM provider

cannot add command 'agent' error (< v1.12.2)

npm install -g careervivid@latest

Job URL is broken / agent says it can't confirm the link The agent uses a live URL verification harness. If a link fails verification it will tell you instead of sending you to a broken page. Search for the role directly on the company's careers page or LinkedIn.

Article shows as public even though I said private Update to v1.12.6+ — articles are now private by default. Use --public or tell the agent "publish publicly" to share with the community.

Permission denied on ~/.careervividrc.json

chmod 600 ~/.careervividrc.json

Resources

• 🌐 careervivid.app
• 🔑 Developer Settings & API Key
• 💳 Pricing & AI Credits
• 🐛 Report an Issue
• 📦 npm Package

MIT License © CareerVivid