pplx-cli

A powerful CLI tool for the Perplexity AI API - search the web and chat with AI from your terminal.

Features

• 🔍 Web Search - Search the web with advanced filtering options
• 💬 AI Chat - Chat with Perplexity's AI models with web-grounded responses
• 🌊 Streaming - Real-time streaming responses
• 📚 Citations - See sources for AI-generated responses
• 🎯 Multiple Models - Support for Sonar, Sonar Pro, and Reasoning models
• ⚙️ Easy Configuration - Simple API key management

Installation

# Using npm
npm install -g pplx-cli

# Using bun
bun install -g pplx-cli

# Using yarn
yarn global add pplx-cli

Quick Start

1. Get your API key from Perplexity API Portal

2. Configure the CLI:

pplx config set-key pplx-xxxxxxxxxxxxxxxx

3. Start asking questions:

pplx ask "What are the latest developments in AI?"

Commands

pplx config

Manage your configuration.

# Save your API key
pplx config set-key <api-key>

# Show current configuration
pplx config show

# Clear configuration
pplx config clear

pplx ask

Ask a single question to Perplexity AI.

# Basic usage
pplx ask "What is the weather in Tokyo?"

# Use a specific model
pplx ask "Explain quantum computing" --model sonar-pro

# Stream the response in real-time
pplx ask "Latest tech news" --stream

# Filter by recency
pplx ask "Breaking news" --recency hour

# Filter by domain
pplx ask "Climate research" --domain nature.com --domain science.org

# Output as JSON
pplx ask "Hello world" --json

# Hide citations
pplx ask "Quick question" --no-citations

# Custom system prompt
pplx ask "Summarize this topic" --system "You are a concise summarizer"

Options:

• -m, --model <model> - Model to use (sonar, sonar-pro, sonar-reasoning, sonar-reasoning-pro)
• -s, --stream - Stream the response in real-time
• -j, --json - Output raw JSON response
• -c, --citations - Show source citations (default: true)
• --no-citations - Hide source citations
• -r, --recency <recency> - Filter by recency (hour, day, week, month, year)
• -d, --domain <domains...> - Filter by domain(s)
• --system <prompt> - Custom system prompt

pplx search

Search the web using Perplexity.

# Basic search
pplx search "climate change research"

# Academic search
pplx search "machine learning papers" --mode academic

# Limit results
pplx search "tech news" --limit 5

# Filter by recency
pplx search "breaking news" --recency day

# Filter by domain
pplx search "AI news" --domain techcrunch.com

# Output as JSON
pplx search "query" --json

Options:

• -l, --limit <number> - Maximum number of results (default: 10)
• -m, --mode <mode> - Search mode (web, academic, sec)
• -r, --recency <recency> - Filter by recency (hour, day, week, month, year)
• -d, --domain <domains...> - Filter by domain(s)
• -j, --json - Output raw JSON response

pplx chat

Start an interactive chat session.

# Start chat
pplx chat

# Use a specific model
pplx chat --model sonar-pro

# With custom system prompt
pplx chat --system "You are a helpful coding assistant"

In-chat commands:

• /clear - Clear conversation history
• /exit or /quit - Exit the chat
• /help - Show help

Options:

• -m, --model <model> - Model to use
• --system <prompt> - Custom system prompt
• -r, --recency <recency> - Filter by recency
• -d, --domain <domains...> - Filter by domain(s)

Available Models

| Model | Description | |-------|-------------| | sonar | Fast, cost-effective model (default) | | sonar-pro | Advanced model with better quality | | sonar-reasoning | Model with reasoning capabilities | | sonar-reasoning-pro | Advanced reasoning model |

Environment Variables

• PERPLEXITY_API_KEY - Your Perplexity API key (alternative to config file)

The config file takes precedence over the environment variable.

Configuration File

The configuration is stored at ~/.config/pplx-cli/config.json.

License

MIT

Links

• Perplexity API Documentation
• GitHub Repository
• Report Issues