This repository contains the Requesty CLI, a command-line AI workflow tool that connects to multiple AI models, enabling model comparison, interactive chat, and intelligent PDF document analysis.
With the Requesty CLI you can:
- Compare AI models side-by-side with streaming responses and real-time performance metrics
- Chat interactively with your favorite AI models in a ChatGPT-style terminal experience
- Analyze PDF documents using AI to extract insights and answer questions about your documents
- Access 100+ AI models from OpenAI, Anthropic, Google, Mistral, xAI, and more
- Stream responses in real-time with beautiful terminal formatting
- Track usage and costs with built-in analytics and feedback systems
Prerequisites: Ensure you have Node.js version 18 or higher installed.
Execute the following command in your terminal:
npm install -g requesty-cli
requestyOr run directly with npx:
npx requesty-cli- Sign up at Requesty.ai for a free account
- Generate an API key from your dashboard
- Set it as an environment variable:
export REQUESTY_API_KEY="your-api-key-here"You are now ready to use the Requesty CLI!
Once the CLI is running, you can start interacting with AI models from your terminal.
Test multiple models with a single prompt:
requesty
> 🚀 Quick Start (5 default models)
# Or compare specific models
requesty quick-start "Explain quantum computing"true"gpt-4o,claude-3-5-sonnet"Start a ChatGPT-style conversation:
requesty chat
> 💬 Regular Chat (ChatGPT-style)
> Choose your model: Claude Sonnet 4
> 💬 You: Help me write a Python web scraper
> 🤖 AI: I'll help you create a Python web scraper...Or start directly with a specific model:
requesty chat openai/gpt-4o
requesty chat anthropic/claude-sonnet-4-20250514Chat with your PDF documents:
requesty pdf-chat research-paper.pdf
> 🔍 Your first question about the PDF: What are the main findings?> 🤖 Assistant: Based on the document, the main findings are...- Side-by-side testing of multiple AI models
- Streaming responses with real-time output
- Performance metrics including response time, token usage, and cost
- Customizable model selection from 100+ available models
- Continuous conversations with context retention
- Featured models with smart categorization
- Recent models sorted by creation date
- Built-in commands:
help,info,summary,clear,exit - Feedback system to improve response quality
- Intelligent document parsing with markdown conversion
- Context-aware responses based on document content
- Multi-turn conversations about your documents
- Support for complex PDFs including technical papers and reports
- Secure API key storage with encryption
- Rate limiting and request management
- Response caching for improved performance
- Comprehensive error handling and recovery
requesty # Start interactive menurequesty chat [model] # Start chat session
requesty pdf-chat <file># Analyze PDF document
requesty security # Check security status
requesty --help # Show help information# Global options
-k, --api-key <key># API key for authentication
-t, --timeout <ms># Request timeout (default: 60000)
--temperature <temp># Response temperature (default: 0.7)# Chat options
requesty chat --temperature 0.9 --timeout 30000
# PDF chat options
requesty pdf-chat document.pdf --model openai/gpt-4oREQUESTY_API_KEY="your-api-key"# Your Requesty API key
DEBUG=true # Enable debug loggingThe CLI supports models from multiple providers:
- OpenAI: GPT-4o, GPT-4.1 Turbo, GPT-4 Mini
- Anthropic: Claude Sonnet 4, Claude Haiku 4
- Google: Gemini 2.5 Flash, Gemini 2.0 Pro
- Mistral: Mistral Large, Mixtral
- xAI: Grok 2.5
- DeepSeek: DeepSeek V3
- And many more...
> Help me refactor this functionto use async/await
> Write unit tests for the authentication module
> Explain this regex pattern: /^(?=.*[a-z])(?=.*[A-Z])(?=.*\d)[a-zA-Z\d]{8,}$/requesty pdf-chat specification.pdf
> Summarize the technical requirements
> What are the API endpoints described?> Generate implementation notes based on section 3> Explain the differences between TCP and UDP
> Show me examples of Python decorators
> How do I implement a binary search tree?> Write a professional email declining a meeting
> Create a README template for my project
> Generate test data for a user databasenpm install -g requesty-clinpm install --save-dev requesty-cligit clone https://github.com/requestyai/requesty-cli
cd requesty-cli
npm install
npm run build
npm link# Check if key is setecho$REQUESTY_API_KEY# Set key for current sessionexport REQUESTY_API_KEY="your-key"# Set key permanently (add to ~/.bashrc or ~/.zshrc)echo'export REQUESTY_API_KEY="your-key"'>>~/.bashrc- Ensure you have an active internet connection
- Check if you're behind a corporate firewall
- Verify the API endpoint is accessible:
curl https://router.requesty.ai/v1/models
- Use streaming mode for faster perceived responses
- Select appropriate models for your use case (smaller models for simple tasks)
- Enable response caching for repeated queries
We welcome contributions! Please see our Contributing Guide for details.
# Install dependencies
npm install
# Run tests
npm test# Build the project
npm run build
# Run in development mode
npm run devThis project is licensed under the MIT License - see the LICENSE file for details.
- Website: requesty.ai
- Documentation: docs.requesty.ai
- Support: support@requesty.ai
Built with ❤️ by the Requesty team. Special thanks to all contributors and the open-source community.
Note: Requesty CLI is in active development. Features and commands may change. Always refer to requesty --help for the most up-to-date information.