pplx

pplx is a command-line client that lets you chat with or query the Perplexity AI API. The project provides an unofficial CLI interface to interact with Perplexity, including an interactive chat mode and a query mode with various options for controlling model, search scope, and response styling. You can enable shell completion to speed up working with commands, subcommands, and flags, and you can use the built-in completion utilities to easily discover models, search modes, recency filters, and other customizable parameters. Typical use cases include performing a quick question-and-answer session with the AI, or executing more advanced queries with explicit model selection, domain restrictions, and response controls.

Prerequisites:

• Docker installed and running (for the recommended container-based run).
• Optional: Access to a container registry if you prefer building your own image.

Installation steps (Docker, recommended):

1. Ensure Docker is installed and running on your machine.

2. Run the MCP server via the provided container image: docker run -i sgaunet/pplx

Alternative: If you want to run locally without Docker and there is a binary available for your platform, you would typically:

• Go to the project repository and build the binary: go install github.com/sgaunet/pplx@latest
• Run the binary directly: pplx [command]

Note: The exact binary name or installation command may differ if the project releases prebuilt binaries for your OS. Refer to the official release page for binaries if you prefer not to build from source.

Tips and potential issues:

• If you’re using the Docker image, ensure you have network access for Perplexity API calls and that any required API keys or authentication are provided via environment variables as documented by the container image or the project. Common env vars may include API keys or tokens; placeholder values can be used during testing.
• The CLI supports commands such as chat, query, and version. Use pplx [command] --help to learn more about each command’s options.
• Shell completion can significantly speed up command discovery. If completions aren’t loaded, follow the Troubleshooting guidance in the README of the project’s completion feature.
• When using query, you can control model selection, search domains, recency, image handling, and generation parameters to tailor results.
• If you encounter permission errors when installing system-wide completions or binaries, prefer running with appropriate privileges or using user-local installation paths.