Skip to main content

AI-powered prompt engineering CLI with multi-agent system

Project description

Promptify

Install from PYPI Promptify Banner

Transform vague ideas into professional engineering specifications using AI agents.

PyPI version Python 3.12 License: MIT


Introduction

Promptify is an advanced AI-powered CLI tool designed to bridge the gap between abstract user intent and concrete technical requirements. It employs a multi-agent architecture (Triage, Critic, Expert, and Prompt Smith) to iteratively refine, critique, and enhance your prompts, delivering production-ready specifications for software projects.

No more back-and-forth. Just Promptify and build.

Features

  • Multi-Agent Architecture:
    • Triage Agent: Understands intent and routes to the right expert.
    • Critic Agent: Identifies gaps, ambiguities, and missing context.
    • Expert Agent: Provides domain-specific architectural advice.
    • Prompt Smith: Synthesizes everything into a perfect final prompt.
  • Dynamic Model Support: Switch seamlessly between Cerebras (fast/free), OpenAI (GPT-4), Anthropic (Claude 3.5), or Local LLMs.
  • Interactive CLI & TUI: Beautiful terminal user interface built with Textual and Rich.
  • Flexible Configuration: Easy YAML-based configuration with environment variable support (.env).

Installation

Install easily via PyPI:

pip install pfy

Usage

1. Quick Start

Run the interactive mode to type your query and see the agents in action:

promptify refine

Or with a direct query:

promptify refine "Build a flappy bird game in python"

2. Configuration

Manage your LLM provider settings easily with the interactive TUI:

promptify config

Or via command line arguments:

# Switch to OpenAI GPT-4
promptify config --provider openai --model gpt-4

# Switch to free Cerebras tier (Default)
promptify config --provider cerebras --model cerebras/llama3.1-8b

# View current config
promptify config --show

3. Advanced Usage

Save the refined spec to a file or change format:

# Save to a markdown file
promptify refine "Create a CRM system" --output specs.md

# Output plain JSON for piping to other tools
promptify refine "Fix my regex" --format json

Setup API Keys

Promptify works best when you provide your own API keys. You can set them via the promptify config TUI or by setting environment variables in your shell (or a .env file).

export CEREBRAS_API_KEY=your_key_here
export OPENAI_API_KEY=sk-...
export ANTHROPIC_API_KEY=sk-...

Architecture

graph TD
    User[User Input] --> T[Triage Agent]
    T --> C[Critic Agent]
    C --> E[Expert Agent]
    E --> S[Prompt Smith]
    S --> Final[Refined Artifact]
    
    subgraph Core
        Provider[Model Provider]
        Config[Configuration]
    end
    
    T -.-> Provider
    C -.-> Provider
    E -.-> Provider
    S -.-> Provider

Contributing

Contributions are welcome! Please visit our GitHub Repository to report issues or submit PRs.

License

Distributed under the MIT License. See LICENSE for more information.


Built with ❤️ by Siva Netizen

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

pfy-0.1.6.tar.gz (20.2 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

pfy-0.1.6-py3-none-any.whl (27.2 kB view details)

Uploaded Python 3

File details

Details for the file pfy-0.1.6.tar.gz.

File metadata

  • Download URL: pfy-0.1.6.tar.gz
  • Upload date:
  • Size: 20.2 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for pfy-0.1.6.tar.gz
Algorithm Hash digest
SHA256 367d86254d8310f35c5823d1135d2f5406214a28230c58b4043eb7434f3204cf
MD5 63130163c419855ad3b5307675ef77aa
BLAKE2b-256 f243d8f044c80c262c8d14c732780beb978dd5b1b5561fcf94fc6320acc11e0e

See more details on using hashes here.

File details

Details for the file pfy-0.1.6-py3-none-any.whl.

File metadata

  • Download URL: pfy-0.1.6-py3-none-any.whl
  • Upload date:
  • Size: 27.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.3

File hashes

Hashes for pfy-0.1.6-py3-none-any.whl
Algorithm Hash digest
SHA256 10a9a17f5159470641ae38a7e7deb0c5d1b51f7885b1c82cc8ca37e344120156
MD5 5fd2903db7d517e379182e1387504e1f
BLAKE2b-256 3e5a3114c4dbb4ad5f6d0546d62bdd7622be611c4a969f4b098c9bb3c896e0a2

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page