GitHub Docs
Demos

Demo 1: First Run & Setup

Overview

This demo walks you through installing VibeCody, configuring your first AI provider, and sending your first chat message. By the end, you will have a working VibeCLI installation connected to an AI provider of your choice.

Time to complete: ~5 minutes

Prerequisites

  • macOS, Linux, or Windows (WSL2)
  • Rust toolchain (1.75+) if building from source, or Docker
  • An API key for at least one AI provider (Ollama works offline without a key)

Step-by-Step Walkthrough

Step 1: Install VibeCody

Choose one of three installation methods.

Option A: One-liner installer (recommended)

curl -fsSL https://raw.githubusercontent.com/TuringWorks/vibecody/main/install.sh | bash

The installer downloads a pre-built binary for your platform and verifies the SHA-256 checksum before placing it in your PATH.

Option B: Build from source with Cargo

cargo install --git https://github.com/TuringWorks/vibecody.git vibecli

Or clone and build locally:

git clone https://github.com/TuringWorks/vibecody.git
cd vibecody
cargo build --release -p vibecli
# Binary is at target/release/vibecli

Option C: Docker

docker pull ghcr.io/turingworks/vibecody:latest
docker run -it --rm \
  -e ANTHROPIC_API_KEY="$ANTHROPIC_API_KEY" \
  ghcr.io/turingworks/vibecody:latest "Hello"

For air-gapped environments with a local Ollama sidecar:

docker compose up -d

Step 2: Verify the installation

vibecli --version

Expected output:

vibecli 0.5.1

Step 3: Set up API keys

You can provide API keys via environment variables or the config file.

Environment variables (quick start):

# Pick one (or more) provider
export ANTHROPIC_API_KEY="sk-ant-..."
export OPENAI_API_KEY="sk-..."
export GEMINI_API_KEY="AIza..."
export GROK_API_KEY="xai-..."

Config file (persistent):

VibeCLI stores its configuration at ~/.vibecli/config.toml. Create or edit it:

mkdir -p ~/.vibecli
cat > ~/.vibecli/config.toml << 'EOF'
[claude]
enabled = true
api_key = "sk-ant-..."
model = "claude-sonnet-4-6"

[openai]
enabled = true
api_key = "sk-..."
model = "gpt-4o"
EOF

Using Ollama (fully offline, no key needed):

# Install Ollama from https://ollama.com
ollama pull llama3

Then set Ollama as the default:

cat > ~/.vibecli/config.toml << 'EOF'
[ollama]
enabled = true
model = "llama3"
EOF

Step 4: Send your first message

vibecli chat "Hello! What can you help me with?"

You should see a streaming response from your configured AI provider.

Step 5: Configure the default provider

Switch your default provider at any time:

# Via CLI flag
vibecli --provider openai "Explain quicksort"

# Via REPL (just run vibecli with no arguments to enter the REPL)
vibecli
> /model codellama

Step 6: Explore help

View all available commands:

vibecli --help

AI-powered coding assistant for the terminal

Usage: vibecli [OPTIONS] [MESSAGE]...

Arguments:
  [MESSAGE]...  One-shot chat message

Options:
  -p, --provider <PROVIDER>  AI provider [default: ollama]
  -m, --model <MODEL>        Model name
      --tui                  Launch terminal UI
      --agent <TASK>         Run agent on a task
      --exec <TASK>          CI/non-interactive mode
      --serve                Start HTTP daemon
      --review               Run code review
      --doctor               Health check
      --help                 Print help
      --version              Print version

Inside the REPL, type /help for interactive commands:

vibecli
> /help

Available REPL Commands:
  /help           Show this help message
  /model          Switch model
  /clear          Clear conversation
  /history        Show conversation history
  /save           Save conversation
  /load           Load conversation
  /tools          List available tools
  /hooks          Manage agent hooks
  /orchestrate    Workflow orchestration
  /sessions       Session management
  /quit           Exit REPL

Step 7: Start the HTTP daemon (optional)

Run VibeCLI as a persistent background service:

vibecli --serve --port 7878 --provider ollama

Test the health endpoint:

curl http://localhost:7878/health

{
  "status": "ok",
  "version": "0.5.1",
  "provider": "ollama"
}

VibeUI Setup

If you want the desktop IDE experience:

cd vibeui
npm install
npm run tauri dev

VibeUI shares the same ~/.vibecli/config.toml configuration, so any provider setup you did above carries over.

Demo Recording

The following JSON recording can be used for automated demo playback with VibeCody’s built-in demo runner.

{
  "meta": {
    "title": "First Run & Setup",
    "description": "Install VibeCody, configure an AI provider, and send your first chat message.",
    "duration_seconds": 120,
    "version": "1.0.0"
  },
  "steps": [
    {
      "id": 1,
      "action": "shell",
      "command": "vibecli --version",
      "description": "Verify VibeCLI is installed",
      "expected_output_contains": "vibecli",
      "delay_ms": 1000
    },
    {
      "id": 2,
      "action": "shell",
      "command": "mkdir -p ~/.vibecli",
      "description": "Create config directory",
      "delay_ms": 500
    },
    {
      "id": 3,
      "action": "write_file",
      "path": "~/.vibecli/config.toml",
      "content": "[ollama]\nenabled = true\nmodel = \"llama3\"\n",
      "description": "Write default provider config",
      "delay_ms": 500
    },
    {
      "id": 4,
      "action": "shell",
      "command": "vibecli chat \"Hello! What can you help me with?\"",
      "description": "Send first chat message",
      "expected_output_contains": "help",
      "delay_ms": 5000,
      "typing_speed_ms": 50
    },
    {
      "id": 5,
      "action": "shell",
      "command": "vibecli --help",
      "description": "Explore available commands",
      "expected_output_contains": "Options",
      "delay_ms": 2000
    },
    {
      "id": 6,
      "action": "repl",
      "commands": [
        { "input": "/help", "delay_ms": 2000 },
        { "input": "What is the capital of France?", "delay_ms": 4000 },
        { "input": "/quit", "delay_ms": 500 }
      ],
      "description": "Interactive REPL session exploring help and sending a message"
    },
    {
      "id": 7,
      "action": "shell",
      "command": "vibecli --serve --port 7878 --provider ollama &",
      "description": "Start HTTP daemon in background",
      "delay_ms": 2000
    },
    {
      "id": 8,
      "action": "shell",
      "command": "curl -s http://localhost:7878/health | python3 -m json.tool",
      "description": "Verify daemon health endpoint",
      "expected_output_contains": "\"status\": \"ok\"",
      "delay_ms": 1500
    }
  ]
}

Troubleshooting

Problem Solution
vibecli: command not found Add the install directory to your PATH: export PATH="$HOME/.cargo/bin:$PATH"
Provider error: missing API key Set the appropriate environment variable or add it to ~/.vibecli/config.toml
Connection refused on Ollama Make sure Ollama is running: ollama serve
Docker permission denied Run with sudo or add your user to the docker group

What’s Next