Skip to content

danaia/pairdCLI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

22 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Paird CLI

AI-powered spec-driven development - Build apps through conversation with your local LLM.

Python 3.9+ License: MIT

What is Paird CLI?

Paird CLI is a command-line tool that helps you build software through a structured, AI-assisted workflow. It takes your idea and generates:

  1. Requirements - What to build (user stories, acceptance criteria)
  2. Design - How to build it (architecture, data models)
  3. Tasks - Step-by-step implementation plan
  4. Code - Actual working code files
  5. Tests - Verification tests

The magic: Use two terminals - one generates code, one lets you chat and iterate in real-time.

Quick Demo

# Terminal 1: Generate a project
paird-cli "create a todo app with dark theme" --autopilot --keep-alive

# Terminal 2: Watch and iterate
paird-cli --watch

# In the watch terminal, refine your app:
> /generate add a delete button to each todo
> /spec add user authentication with Google login

Installation

# Clone and install
git clone https://github.com/danaia/pairdCLI.git
cd pairdCLI
pip install -e .

# Configure your LLM (LM Studio, Ollama, etc.)
paird-cli config init
paird-cli config set llm.endpoint http://localhost:1234/v1
paird-cli config set llm.model qwen3-30b-instruct

Usage Modes

🚀 Quick Start (Single Terminal)

paird-cli "build a REST API with Express"

⚡ Autopilot Mode (Hands-off)

paird-cli "create a landing page" --autopilot --no-tests

🔥 Dual Terminal Mode (Recommended)

Terminal 1 - Work:

paird-cli "create a modern login page" --autopilot --keep-alive

Terminal 2 - Watch & Iterate:

paird-cli --watch

# Then use commands like:
> /generate make background dark
> /spec add forgot password flow

Key Features

Feature Description
Dual Terminal Mode One terminal generates, one lets you iterate
/spec Command Generate detailed specs with iterative refinement
/generate Command Quick code changes in real-time
Auto npm install Dependencies installed automatically
Resume Support Save progress, resume anytime
Local-first Works with LM Studio, Ollama, any OpenAI-compatible API

Watch Terminal Commands

Command Description
/generate <prompt> Quick code generation
/spec <prompt> Detailed spec with refinement loop
/read <file> Add file to context
/status Show pipeline status
/help All commands

Configuration

paird-cli config init                           # Create config
paird-cli config set llm.endpoint <url>         # Set LLM endpoint
paird-cli config set llm.model <model-name>     # Set model
paird-cli config show                           # View config

Recommended: Qwen3 30B+ models for best results (60K context window)

Project Structure

your-project/
├── .paird/specs/project-name/
│   ├── requirements.md    # Generated requirements
│   ├── design.md          # Technical design
│   ├── tasks.md           # Implementation tasks
│   └── session.json       # Resume state
├── src/                   # Generated code
├── tests/                 # Generated tests
└── package.json           # Auto-generated

Documentation

Requirements

  • Python 3.9+
  • LM Studio, Ollama, or OpenAI-compatible endpoint
  • macOS or Linux

Contributing

Contributions welcome! Please open an issue or PR.

License

MIT License - see LICENSE for details.


Built with ❤️ using Rich, Prompt Toolkit, and local LLMs.

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors