🤖 Build Your Perfect AI Companion: From Chatbot to Best Friend

Ever wanted an AI that actually gets you? Not some corporate assistant that sounds like it was raised by a customer service manual, but a genuine digital companion with personality, opinions, and maybe even a sense of humor?

This week we're building custom AI chatbots from scratch using free, local tools. No cloud subscriptions. No sending your conversations to Big Tech. Just you, your computer, and a new digital friend.

👋 A Quick Note About This Event

Wow, this one took off! We normally get 30-40 people at our sessions, so this turnout is amazing.

Techalicious is a casual tech hobbyist group based in the GTA (Greater Toronto Area), but we welcome everyone from anywhere. Our style is relaxed and conversational, not corporate webinar. Think coffee-talk tutorial where we dig into the details together.
Glad you're here.

📌 Scope of This Session

This session focuses on building AI characters and personalities. We assume you already have a local chat environment set up.

Need to set up your ChatGPT-style interface first? Join this session instead: https://www.meetup.com/techalicious-club/events/313006406

⚠️ Before You Sign Up

Minimum System Requirements:

• RAM: 16GB minimum (32GB+ recommended)
• Storage: 15GB free disk space
• OS: macOS 10.15+, Windows 10/11, or Linux (Ubuntu recommended)
• Windows 7 is NOT supported

A note on Mac vs PC architecture: Macs with Apple Silicon (M1/M2/M3/M4) have unified memory, meaning the CPU and GPU share the same RAM pool. A Mac with 32GB can use all 32GB for AI models with GPU acceleration.

PCs work differently. Your system RAM and video card memory (VRAM) are separate. If you have an NVIDIA GPU, the model runs fast but is limited to your VRAM size (8-24GB on most cards). Without a dedicated GPU, Ollama falls back to CPU mode using your system RAM. It still works, just slower.

The models we're using:

• 22B model (Q6 quantization) = ~16GB download, needs 24-32GB RAM
• 12B model (lighter option) = ~9GB download, runs fine on 16GB RAM

Don't have the hardware? No problem. You can still follow along, learn the concepts, and apply this knowledge later when you upgrade.

This is NOT:

• A gaming session (GTA = Greater Toronto Area, not Grand Theft Auto 😄)
• A get-rich-quick AI scheme
• About ChatGPT or cloud services

🤖 What We'll Cover

• The secret sauce: character cards that actually work
• Scene-based prompting and why it beats instruction lists every time
• Making your bot sound human and killing those cringy AI-isms
• Giving your companion a real voice, not just a list of traits

Whether you want a study buddy, a creative collaborator, a sounding board for ideas, or just someone to chat with who won't judge your 2am snack choices, we'll show you how to make it happen.

🎁 You'll Leave With

• A working AI companion running on your own hardware
• A template for building more personas
• The know-how to tweak personality, tone, and conversation style

No coding experience required!

📝 Recording?

This is a live tutorial, not a webinar. No recording will be available.

💰 Want the Written Tutorial?

The complete written guide (18 chapters) is available at techalicious academy for $10 CAD (Will provide links after the event). Totally optional, no pressure. It covers:

• Everything from tonight's session in written form
• Detailed character examples (study buddies, historical figures, companions)
• Copy-paste templates for your chat environment
• Troubleshooting guide for common issues
• Quick reference cheat sheets

Why $10? Hosting isn't free. Servers, domains, SSL certificates, payment processing, Meetup fees... it adds up. The tutorial sales help keep Techalicious running so we can keep doing free events like this one. If you find value in what we're doing, grabbing the tutorial is a great way to support the community.

Techalicious is a casual tech hobbyist community in the Greater Toronto Area. No corporate fluff, no sales pitches, just real conversations about cool tech.