#4 AI Use-Cases with different LLM models

Ollama AI Daily Toolkit — Ready-to-Run YAML + Prompts

Paste this entire HTML into Blogger's HTML view. Each code block has Prism syntax highlighting and a Copy button for easy copy-&-paste into your terminal or editor.

Quick Instructions

1) Click Copy on any code block and paste into your machine.
2) Use ollama run -f <yaml-file> in terminal after downloading or placing models. 3) Replace model names with locally available/lightweight alternatives if needed.

1) daily_planner.yaml — Smart Personal Assistant & Daily Planner

Models: mistral (planner), phi3 (summarizer), codgemma (automator)

models:
  - name: mistral
    role: planner
  - name: phi3
    role: summarizer
  - name: codgemma
    role: automator
workflow:
  - mistral -> phi3 -> codgemma

Prompt Example

Plan my Monday schedule with meetings, travel, and 2 hours of deep work.
Summarize unread emails and add tasks from them.

Expected Output

🗓️ Monday Plan:
- 9:00–10:00 AM: Weekly Meeting (Zoom)
- 10:30–11:00 AM: Travel to Office
- 11:30–1:30 PM: Deep Work – Project Draft
- 2:00 PM: Review team emails → Add 2 tasks (Follow-up on client proposal, update budget sheet)

2) home_energy_optimizer.yaml — Home Energy Optimizer

Models: llama3 (advisor), tinyllama (data_analyzer), phi3 (predictor)

models:
  - name: llama3
    role: advisor
  - name: tinyllama
    role: data_analyzer
  - name: phi3
    role: predictor
workflow:
  - tinyllama -> phi3 -> llama3

Prompt Example

Analyze my last 7 days of electricity usage data.
Suggest when to run washing machine and water heater to save power.

Expected Output

⚡ Power Optimization Plan:
- Washing machine: Run between 2–4 PM (solar surplus)
- Water heater: 7–8 AM, avoid 6–8 PM (peak tariff)
Estimated savings: ₹320/month.

3) travel_buddy.yaml — Multilingual Personal Translator & Travel Buddy

Models: mistral (translator), phi3 (culture_advisor), whisper (speech_to_text)

models:
  - name: mistral
    role: translator
  - name: phi3
    role: culture_advisor
  - name: whisper
    role: speech_to_text
workflow:
  - whisper -> mistral -> phi3

Prompt Example

Translate this sentence to Japanese: "Where is the nearest train station?"
Also tell me if it's polite enough.

Expected Output

Japanese: 一番近い駅はどこですか？ (Ichiban chikai eki wa doko desu ka?)
✅ This phrase is polite and suitable for speaking with strangers.

4) budget_assistant.yaml — Smart Budget & Shopping Assistant

Models: phi3 (categorizer), mistral (extractor), llama3 (optimizer)

models:
  - name: phi3
    role: categorizer
  - name: mistral
    role: extractor
  - name: llama3
    role: optimizer
workflow:
  - mistral -> phi3 -> llama3

Prompt Example

Here are my last week’s expenses:
- Groceries ₹1200
- Snacks ₹350
- Taxi ₹500
- Coffee ₹300
Suggest where I can save 10%.

Expected Output

💰 Expense Optimization:
- Reduce snacks & coffee: Target ₹400/week (save ₹250)
- Try monthly grocery packs: Estimated ₹150 saving.
✅ Total potential saving: ₹400 (≈11%)

5) study_coach.yaml — Local Study & Skill Coach

Models: llama3 (tutor), phi3 (quizmaster), codgemma (code_assistant)

models:
  - name: llama3
    role: tutor
  - name: phi3
    role: quizmaster
  - name: codgemma
    role: code_assistant
workflow:
  - llama3 -> phi3 -> codgemma

Prompt Example

Teach me Python basics in one week.
Give me today’s topic, one example, and a 2-question quiz.

Expected Output

📘 Python Week 1 – Day 1: Introduction to Variables
Example:
x = 5
name = "Ravi"
print(name, x)

🧩 Quiz:
1. What is a variable?
2. How do you print multiple items in Python?

Bonus) ai_daily_toolkit.yaml — All-in-One Launcher (Menu Mode)

A simple menu-style YAML referencing all assistant configs.

models:
  - name: llama3
  - name: mistral
  - name: phi3
  - name: codgemma
  - name: tinyllama
  - name: whisper
menu:
  - Smart Planner: daily_planner.yaml
  - Home Optimizer: home_energy_optimizer.yaml
  - Travel Buddy: travel_buddy.yaml
  - Budget Assistant: budget_assistant.yaml
  - Study Coach: study_coach.yaml

Run command

ollama run -f ai_daily_toolkit.yaml

README: Usage Tips & Model Notes

1. Download models: Replace model names with ones you have or download lightweight alternatives in Ollama first.
2. Permissions: If Blogger strips external scripts, add Prism and this script via a trusted host or inline them (but Blogger often allows CDNs).
3. Customise: Edit prompts to match your locale (currency, time format) or swap model names for your available local models.

Made for quick copy-and-run demos — tweak models & prompts to suit your environment. 👍