User Guide

This comprehensive user guide covers everything you need to know to use Perspt effectively, from basic conversations to advanced productivity techniques.

User Guide Contents

Overview

Perspt is a high-performance terminal-based chat application built with Rust that provides a unified interface for interacting with multiple Large Language Model (LLM) providers. Using the modern genai crate (v0.3.5), Perspt offers seamless access to the latest AI models with enhanced streaming capabilities, robust error handling, and intuitive CLI functionality.

Key Features:

  • Multi-Provider Support: OpenAI, Anthropic, Google, Groq, Cohere, XAI, and Ollama

  • Latest Models: Support for reasoning models (o1-mini, o1-preview, o3-mini), GPT-4.1, Claude 3.5, Gemini 2.5 Pro

  • Real-time Streaming: Enhanced streaming with proper reasoning chunk handling

  • Robust CLI: Working command-line arguments with model validation

  • Beautiful UI: Responsive terminal interface with markdown rendering

🚀 Basic Usage

Learn the fundamentals of chatting with AI models, keyboard shortcuts, and everyday usage patterns.

Basic Usage
⚡ Advanced Features

Discover powerful features like input queuing, markdown rendering, and productivity workflows.

Advanced Features
🔀 Provider Guide

Complete guide to all supported AI providers, their models, and specific configuration options.

AI Providers
🛠️ Troubleshooting

Solutions to common issues, error messages, and optimization tips.

Troubleshooting

Quick Reference

Essential Keyboard Shortcuts

Shortcut

Action

Enter

Send message

Ctrl+C

Exit application

↑/↓ Keys

Scroll chat history

Page Up/Down

Fast scroll

Ctrl+L

Clear screen

Common Commands

# Start with default settings (gpt-4o-mini)
perspt

# Use specific model with validation
perspt --model gpt-4.1

# Switch provider and model
perspt --provider-type anthropic --model claude-3-5-sonnet-20241022

# List available models for current provider
perspt --list-models

# Use custom configuration file
perspt --config my-config.json

# Override API key from command line
perspt --api-key your-api-key

# Use provider profile from config
perspt --provider work-profile

Typical Workflows

Daily Development

  1. Code Review: Paste code and ask for feedback

  2. Documentation: Generate or improve documentation

  3. Debugging: Discuss error messages and solutions

  4. Learning: Ask about new technologies or concepts

Research and Writing

  1. Information Gathering: Ask questions about topics

  2. Content Creation: Get help with writing and editing

  3. Brainstorming: Generate ideas and explore concepts

  4. Fact Checking: Verify information and get references

Getting the Most from Perspt

Best Practices

  • Be Specific: Clear, detailed questions get better answers

  • Provide Context: Include relevant background information

  • Iterate: Build on previous responses for deeper understanding

  • Experiment: Try different models for different types of tasks

Productivity Tips

  • Use Configuration Files: Set up profiles for different use cases

  • Learn Keyboard Shortcuts: Speed up your workflow

  • Leverage Streaming: Keep typing while AI responds

  • Save Important Conversations: Copy valuable responses

What’s Next?

Choose your path based on your experience level:

New Users: Start with Basic Usage to learn the fundamentals.

Experienced Users: Jump to Advanced Features for productivity techniques.

Multi-Provider Users: Check out AI Providers for provider-specific tips.

Having Issues?: Visit Troubleshooting for solutions.

See also