Build with AI.
$0/month.

Install OpenCode

OpenCode is a terminal-based AI agent that reads your files, understands context, and helps you with any task. It's open source and comes with Zen, a built-in gateway that includes free models. Just select one and start working — no accounts needed.

Installation

OpenCode is installed via npm. Make sure you have Node.js 18+ installed first.

brew install node  # if not installed
npm install -g opencode

# Verify installation
opencode --version

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

sudo apt install nodejs npm  # Ubuntu/Debian
# or: sudo dnf install nodejs npm  # Fedora
npm install -g opencode

# Verify installation
opencode --version

Select Zen as Your Provider (Free)

OpenCode comes with Zen, a built-in model gateway with rotating free models. No API key needed, no credit card, no config files — just pick a model and go.

1. Open your terminal and run opencode
2. You'll see a welcome screen — select Zen as your provider
3. Pick one of the free models (they're labeled as free)
4. That's it — start using it immediately

# Launch OpenCode
opencode

# On the welcome screen:
# 1. Select "Zen" as your provider
# 2. Pick any free model from the list
# 3. Start working — no accounts or API keys needed

Your First Task

Let's make sure everything works. Give OpenCode something fun:

# Navigate to any folder
cd ~/Desktop

# Launch OpenCode and give it a task
opencode "Make me a shopping list for a homemade lasagna dinner for 4 people, including ingredients I'd need and estimated quantities"

OpenCode will generate something like this:

## Lasagna Dinner for 4

### Pasta & Sauce
- Lasagna noodles: 12 sheets (one 16oz box)
- Crushed tomatoes: 1 can (28oz)
- Tomato paste: 1 small can (6oz)
- Olive oil: 2 tablespoons

### Meat & Dairy
- Ground beef: 1 lb
- Ricotta cheese: 15oz container
- Mozzarella cheese: 1 lb (shredded)
- Parmesan cheese: 1/2 cup (grated)
- Eggs: 1 large

### Vegetables & Herbs
- Yellow onion: 1 medium
- Garlic cloves: 4
- Fresh basil: 1 small bunch
- Dried oregano: 1 teaspoon
- Salt and pepper to taste

### Side Dish & Bread
- Garlic bread: 1 loaf
- Mixed salad greens: 1 bag (5oz)
- Salad dressing of choice: 1 bottle

Try These Prompts

Copy any of these and paste them into OpenCode to see what it can do:

Plan a week of healthy dinners for 2 people and create a combined shopping list

Organize my Downloads folder — sort files into subfolders by type (images, documents, videos, etc)

Write a professional email declining a meeting request politely

Create a simple budget tracker that reads a CSV file and shows spending by category

Summarize the key points from this article: [paste any article text here]

Create a Python script that renames all photos in a folder with the date they were taken

Upgrade to OpenRouter

Zen's free models are great for getting started. For more powerful and reliable models, upgrade to OpenRouter — it gives you access to 29+ free models from top providers. No credit card needed.

Sign Up & Get Your API Key

# 1. Go to https://openrouter.ai
# 2. Click "Sign Up" (GitHub, Google, or email)
# 3. Go to Settings → Keys
# 4. Click "Create Key"
# 5. Copy your key — it starts with "sk-or-v1-"

# Your free tier includes credits for free models
# No credit card required to start

Configure OpenCode with OpenRouter

Now point OpenCode at OpenRouter instead of Zen. Paste your API key in place of YOUR_KEY.

Terminal command (creates the file automatically):

mkdir -p ~/.config/opencode && cat > ~/.config/opencode/config.json << 'EOF'
{
  "provider": "openrouter",
  "model": "qwen/qwen3.6-plus:free",
  "api_key": "sk-or-v1-YOUR_KEY",
  "base_url": "https://openrouter.ai/api/v1"
}
EOF

Or edit manually — open ~/.config/opencode/config.json and paste:

{
  "provider": "openrouter",
  "model": "qwen/qwen3.6-plus:free",
  "api_key": "sk-or-v1-YOUR_KEY",
  "base_url": "https://openrouter.ai/api/v1"
}

Or if you prefer a different free model, try one of these alternatives:

qwen/qwen3.6-plus:free        — Best all-round free model
google/gemma-3-27b-it:free     — Google's open model
meta-llama/llama-4-scout:free  — Meta's latest open model

How Free Models Work on OpenRouter

OpenRouter offers models with a :free suffix at no cost. These are provided by model creators as a way to let developers try their models. The free tier is genuinely free — no hidden charges, no credit card required.

Free models may have:

• Rate limits (requests per minute)
• Lower priority during high traffic
• Potential usage caps per day

The free tier allows up to 50 requests per day and 20 requests per minute. For personal development, this is usually enough. If you need more, purchasing 10+ credits raises the limit to 1,000 requests/day.

Add Obsidian as Your Knowledge Base

Supercharge your workflow with Obsidian. It's a powerful local-first note-taking app that stores your notes as plain Markdown files. With AI plugins, it becomes your personal knowledge base that can summarize, search, and generate content using the same free models from OpenRouter.

Install Obsidian

Download Obsidian from obsidian.md — it's free for personal use on all platforms (macOS, Windows, Linux, iOS, Android).

# Download from obsidian.md or use Homebrew
brew install --cask obsidian

# Download the .deb or .AppImage from obsidian.md
# For Ubuntu/Debian:
sudo dpkg -i obsidian_*.deb
# Or use Snap:
sudo snap install obsidian --classic

Set Up the Copilot Plugin

The Copilot plugin brings AI directly into your Obsidian vault. Here's how to set it up:

1. Open Obsidian and go to Settings → Community Plugins
2. Click Browse and search for "Copilot"
3. Install and enable the plugin
4. In Copilot settings, set the provider to OpenRouter
5. Enter your OpenRouter API key
6. Select your preferred free model from OpenRouter (look for models with the :free suffix)

In the Copilot plugin settings, use these values:

{
  "provider": "openrouter",
  "apiKey": "sk-or-v1-YOUR_KEY",
  "model": "qwen/qwen3.6-plus:free",
  "baseUrl": "https://openrouter.ai/api/v1",
  "temperature": 0.7,
  "maxTokens": 4096
}

You can swap qwen/qwen3.6-plus:free for any model from the OpenRouter free models list.

What You Can Do with Obsidian + AI

• Chat with your notes — Ask questions and get answers based on your vault content
• Summarize long documents — One-click summaries of meeting notes, articles, research
• Generate content — Draft blog posts, documentation, or creative writing
• Tag and organize — AI suggests tags and categories for new notes
• Translate — Translate notes between languages instantly
• Extract action items — Pull tasks and todos from meeting notes automatically

Alternative: Obsidian Text Generator Plugin

Another excellent option is the Text Generator plugin, which supports OpenRouter out of the box and offers more granular control over prompts and templates.

AGENTS.md — Teach Your Agent About Your Project

This is the secret weapon that makes AI agents genuinely useful. Create an AGENTS.md file in your project root to give your AI agent context about your project's conventions, architecture, and standards.

Create Your AGENTS.md

Create an AGENTS.md file in your project root to give AI agents context about your project:

# Project: My SaaS App

## Tech Stack
- Next.js 15 (App Router)
- TypeScript
- Tailwind CSS + shadcn/ui
- PostgreSQL + Prisma
- Deployed on Vercel

## Coding Standards
- Use functional components with hooks
- Prefer server components where possible
- All API routes use tRPC
- Tests with Vitest + Testing Library
- ESLint + Prettier enforced

## Architecture
- /app — Next.js app router pages
- /components — Shared UI components
- /lib — Utilities and shared logic
- /server — Server actions and API routes
- /prisma — Database schema and migrations

## When Adding Features
1. Check if a similar component exists
2. Follow the existing file structure
3. Add tests for new functionality
4. Update this file if architecture changes

OpenCode reads this file automatically and uses it to understand your project conventions. This means every code suggestion follows your team's standards.