1.05 – Practical AI Tools | AI Course

🛠️ 1.05 – Practical AI Tools

📑 On this page

💬 Chatbots & Assistants

📌 ChatGPT (OpenAI)

Most popular AI assistant. GPT-4 powers the paid version (ChatGPT Plus). Features: web browsing, file uploads, image generation (DALL-E), code interpreter.

📌 Claude (Anthropic)

Known for strong reasoning, safety, and large context window (200k tokens). Excellent for analyzing long documents.

📌 Gemini (Google)

Multimodal by default – supports text, images, audio, and video. Integrated with Google Workspace.

📌 Perplexity AI

AI-powered search engine with citations. Combines LLM with real-time web search.

💻 AI Coding Tools

📌 GitHub Copilot

AI pair programmer integrated into VS Code, JetBrains, and other IDEs. Autocompletes code based on context and comments.

📌 Cursor

AI-first IDE built on VS Code. Features: Cmd+K for inline editing, chat with codebase, AI terminal commands.

📌 Cline (prev. Claude Dev)

VS Code extension that handles file editing, terminal commands, and browser automation.

📌 Continue

Open-source VS Code extension for AI code assistance. Works with local models or cloud APIs.

🏗️ Development Frameworks

📌 LangChain

Most popular framework for building LLM applications. Provides chains, agents, memory, and integrations with 100+ tools.

📌 LlamaIndex

Specialized framework for RAG (retrieval-augmented generation) applications. Excellent for document indexing and querying.

📌 Semantic Kernel (Microsoft)

Microsoft's lightweight framework for AI orchestration. Integrates well with Azure services.

📌 Ollama

Run open-source LLMs locally (Llama 3, Mistral, Phi-3) on your own hardware.

🔄 Workflow & Agent Tools

📌 n8n (AI-native workflow)

No-code automation platform with AI nodes. Connect LLMs to APIs, databases, and services.

📌 Flowise

Drag-and-drop UI for building LangChain applications. Great for prototyping.

📌 Vector Databases (Quick reference)

📊 Tools Comparison Table

ToolCategoryCostBest for
ChatGPTChatbotFree/$20General conversation
ClaudeChatbotFreeLong documents, reasoning
GitHub CopilotCoding$10/moAutocomplete
CursorCodingFree/$20AI-first IDE
LangChainFrameworkFreeComplex LLM apps
LlamaIndexFrameworkFreeRAG applications
OllamaLocalFreeRun models locally
n8nWorkflowFree/paidAI automations

✍️ Exercises

Exercise 1.05.1 – Choose the right tool

For each scenario, choose the best tool from the list above:

  1. You need an AI to analyze a 500-page legal document.
  2. You're building an internal RAG system for company documents and want an open-source framework specialized for this.
  3. You want autocomplete in your VS Code while writing Python.
  4. You need to process customer documents on-premise for privacy reasons (no cloud APIs).
Answers:
1. Claude (200k context window, good for long documents)
2. LlamaIndex (specialized for RAG)
3. GitHub Copilot or Cursor
4. Ollama (run local open-source models)
Exercise 1.05.2 – Tool research

Pick one tool from the list above, visit its website, and write a one-paragraph summary of its key features and pricing.

Example for Cursor:
Cursor is an AI-first code editor built on VS Code. Key features include: Cmd + K for inline editing (write or modify code with natural language), chat with your entire codebase, AI-predictions for what you'll type next, and integrated AI terminal commands. Pricing: Free tier includes 2 weeks of premium features, then $20/month for Pro (unlimited AI commands, chat with codebase). Cursor uses a mix of GPT-4, Claude 3, and their own models.
Exercise 1.05.3 – Framework exploration

Go to the LangChain documentation website and find one example of a "chain" (e.g., LLMChain). Describe what it does in one sentence.

Example:
An LLMChain is the simplest chain that combines a prompt template with an LLM – you give it user input, it formats the prompt, sends it to the LLM, and returns the response. For example, a chain that summarizes user-provided text.
📌 Tool Selection Tips
📘 Next Module: 1.06 – Build Your First AI App →