Linux command
perplexity 命令
文本
复制后可按需替换文件名、目录或参数。
常用示例
Ask a question
perplexity "What is quantum computing?"
Ask with a specific model
perplexity -m [sonar-pro] "[question]"
Display citations
perplexity -c "[question]"
Show token usage
perplexity -u "[question]"
Search across notes and chats
perplexity rag "[query]"
Set up API key
perplexity setup
说明
perplexity is a command-line interface for Perplexity AI, enabling terminal-based access to their AI-powered search and question-answering capabilities. It leverages Perplexity's Sonar models which combine large language models with real-time web search. The CLI supports multiple models ranging from sonar (fastest) to sonar-reasoning-pro (most capable with deep reasoning). Answers can include citations showing the sources used to generate responses. Advanced features include RAG (Retrieval Augmented Generation) for searching across local notes and chat history using semantic vector search or keyword matching. The knowledge base uses BGE embeddings for lightning-fast semantic search. Chat history and notes are stored locally, enabling export to markdown, JSON, or Excel formats. The tool works offline for local operations while requiring internet connectivity for AI queries.
参数
- -m, --model _model_
- Select language model: sonar, sonar-pro, sonar-reasoning, sonar-reasoning-pro.
- -c, --citations
- Display source citations with the answer.
- -u, --usage
- Show token consumption metrics.
- -v, --verbose
- Enable debug mode with detailed output.
- -g, --glow
- Apply Glow-compatible markdown formatting.
- -a, --api-key _key_
- Provide API key directly instead of environment variable.
FAQ
What is the perplexity command used for?
perplexity is a command-line interface for Perplexity AI, enabling terminal-based access to their AI-powered search and question-answering capabilities. It leverages Perplexity's Sonar models which combine large language models with real-time web search. The CLI supports multiple models ranging from sonar (fastest) to sonar-reasoning-pro (most capable with deep reasoning). Answers can include citations showing the sources used to generate responses. Advanced features include RAG (Retrieval Augmented Generation) for searching across local notes and chat history using semantic vector search or keyword matching. The knowledge base uses BGE embeddings for lightning-fast semantic search. Chat history and notes are stored locally, enabling export to markdown, JSON, or Excel formats. The tool works offline for local operations while requiring internet connectivity for AI queries.
How do I run a basic perplexity example?
Run `perplexity "What is quantum computing?"` in a terminal, then adjust file names, paths, flags, or remote targets for your system.
What does -m, --model _model_ do in perplexity?
Select language model: sonar, sonar-pro, sonar-reasoning, sonar-reasoning-pro.