pls: Lightweight AI CLI Tool

Stop switching contexts. pls is a single Bash script — lightweight, fast, and built for everyday CLI tasks. Seamlessly switch between AI chat and shell commands.

Install

curl -sSL https://raw.githubusercontent.com/cjccjj/pls/main/install.sh | bash

Requires curl and jq. Works on Linux and macOS.

Then export your API key for one of the supported providers:

export OPENAI_API_KEY="sk-..."      # OpenAI (GPT)
export GEMINI_API_KEY="..."         # Google Gemini
export DEEPSEEK_API_KEY="sk-..."    # DeepSeek

Optional: install glow for prettier markdown output.

Usage

pls v0.56

Usage:    pls [messages...]                       # Chat with an input
          > what is llm                           # Continue chat, q or empty input to quit
                                                
Examples: pls                                     # Start without input 
          pls count files                         # ls -1 | wc -l           # shell cmd wait for run
          > include subdirs                       # find . -type f | wc -l  # shell cmd update

Pipe and Chain:          
          echo how to cook rice | pls             # Use input from pipe
          echo rice | pls how to cook             # Args + pipe (task from args, data from pipe)
          pls name a dish | pls -p how to cook    # Chain commands and show piped input with -p

Settings: pls -h                                  # Show this help
          pls edit config                         # config pls and AI model via chat

Config

To switch AI providers, edit ~/.pls/pls.conf and change the profile field (or type edit config in pls chat):

[Global]
profile="deepseek_1"    # change to openai_1, gemini_1, or deepseek_1

[openai_1]
provider="openai"
model="gpt-5-mini"
base_url="https://api.openai.com/v1"
env_key="OPENAI_API_KEY"

[gemini_1]
provider="gemini"
model="gemini-3-flash-preview"
base_url="https://generativelanguage.googleapis.com/v1beta"
env_key="GEMINI_API_KEY"

[deepseek_1]
provider="deepseek"
model="deepseek-v4-flash"
base_url="https://api.deepseek.com/beta"
env_key="DEEPSEEK_API_KEY"

Note: DeepSeek uses the beta endpoint (api.deepseek.com/beta) for reliable structured output via tool calls, avoiding the known JSON mode whitespace bug.

Key Capabilities

• Generate chat messages and shell commands in a single request, it never "thinks twice."
• Switch effortlessly between interactive mode and command mode with shared memory.
• Keep a short-term history, allowing you to pick up anytime without session management.
• Run as a single Bash script, with no extra tools or MCP needed, consumes very few tokens.
• (Experimental) Configure via chat: update yourself, edit config, delete chat history.

Features

• AI chat and shell command generation
• Inline command regeneration and editing
• Supports pipes and command chaining, clean output in piped mode
• Works on Linux and macOS
• Supports OpenAI, Gemini, and DeepSeek APIs

Examples

• You are working on a file, and use pls to help process it:

• You cat the result, are happy with it, and move on something else
• You logout ... and login back
• Then you cd to another working folder where you want to continue the task,
• By simple type pls, the AI picks up the task, and through interative chats you provide new input:

• The AI generated shell commands for you to run,
• And you decided to further tweek the commands,
• Finally, the command runs successfully.

Tips

• Use cat, tail, or any command pipe to feed data into pls.
• You can also feed "knowhow" like my_app --help | pls learn it
• Or generate data for your app pls generate some markdown demo | python my_markdown_render.py
• Give orders in chat mode, refine your request as many times until satisfied.
• Quit chat at any time to run any command you want, and pick it up again anywhere, anytime, by typing pls.
• Use AI for its flexibility to handle anything; for simple commands, run them yourself for efficiency.
• For short context, let AI do it directly; for large datasets, ask AI to generate commands or scripts for you to run.
• If not satisfied with anything, simply provide your feedback.
• You can apply the workflow to tasks like system administration, config file tweaking, log analysis, and script writing and testing.

Update

curl -sSL https://raw.githubusercontent.com/cjccjj/pls/main/install.sh | bash

Or say update yourself in pls chat. Updates preserve your existing config and automatically append any new provider profiles.

Uninstall

sudo rm /usr/local/bin/pls
rm -rf ~/.pls