---

title: Gemini date: 2024-12-09 10:00:00 background: bg-[#4285F4] tags: - AI - Google - Prompts - Multimodal categories: - AI intro: Google Gemini AI assistant cheat sheet covering prompts, features, extensions, and best practices for daily productivity plugins: - copyCode

Overview

Model Versions

Key Capabilities

• Text generation and editing
• Code writing and debugging
• Image understanding and generation
• Document and PDF analysis
• Google Workspace integration
• Real-time web search
• YouTube video summarization

Access Methods

• gemini.google.com - Web interface
• Gemini App - Mobile (iOS/Android)
• Google AI Studio - Developer playground
• Vertex AI - Enterprise API

Prompt Basics {.cols-2}

Prompt Structure

[Context] [Task] [Format] [Constraints]

Example

You are a marketing expert.
Write 5 Instagram caption ideas for a coffee shop.
Use emojis and include hashtags.
Keep each under 150 characters.

Role Assignment

Act as a [role] with expertise in [domain].
Your task is to [specific task].

Output Formatting

Respond in the following format:
- Summary: (2-3 sentences)
- Key Points: (bullet list)
- Recommendation: (1 paragraph)

Google Extensions {.cols-2}

Available Extensions

Using Extensions

Flights Search

@Google Flights Find flights from Sydney to Tokyo
next month, round trip, economy class

Hotel Search

@Google Hotels Find hotels in Melbourne CBD
for 2 nights, under $200/night, with pool

YouTube Summary

@YouTube Summarize this video:
[paste YouTube URL]

Workspace Integration

Gmail Search

@Gmail Find emails from John about the project
proposal from last week

Docs Summary

@Google Docs Summarize the main points from
my document "Q4 Marketing Plan"

Drive Search

@Google Drive Find spreadsheets related to
budget planning from this year

Prompting Techniques

Zero-Shot

Direct question without examples

What are the main differences between
REST and GraphQL APIs?

Few-Shot

Provide examples for pattern learning

Convert to formal language:

Casual: "Hey, can you help me out?"
Formal: "Would you be able to assist me?"

Casual: "That's awesome!"
Formal: "That is excellent."

Casual: "gonna grab some food"
Formal:

Chain of Thought

Step-by-step reasoning

Solve this problem step by step:

A store has 150 items. They sell 30% on Monday,
then receive 50 new items on Tuesday.
How many items do they have now?

Think through each step before answering.

Self-Consistency

Multiple reasoning paths

Analyze this business decision using three
different frameworks:
1. SWOT Analysis
2. Cost-Benefit Analysis
3. Risk Assessment

Then provide a unified recommendation.

Writing Prompts {.cols-2}

Content Creation

Blog Post

Write a blog post about [topic]:
- Target audience: [audience]
- Tone: [casual/professional/technical]
- Length: approximately [X] words
- Include: introduction, main points, conclusion
- Add a compelling headline

Social Media

Create [number] social media posts about [topic]:
- Platform: [Instagram/Twitter/LinkedIn]
- Include relevant hashtags
- Add call-to-action
- Keep within character limits

Email Writing

Professional Email

Write a professional email:
- Purpose: [meeting request/follow-up/inquiry]
- Recipient: [role/relationship]
- Key points to cover: [list]
- Tone: [formal/friendly professional]
- Include subject line

Editing & Rewriting

Improve this text:
- Fix grammar and spelling
- Improve clarity and flow
- Maintain the original meaning
- Make it more engaging

Original: [paste text]

Code Prompts {.cols-2}

Code Generation

Write a [language] function that:
- Input: [description]
- Output: [description]
- Requirements: [list requirements]
- Include error handling
- Add comments explaining the logic

Code Review

Review this code for:
1. Bugs and errors
2. Security vulnerabilities
3. Performance issues
4. Best practices
5. Readability improvements

```[language]
[paste code]

### Debugging

Debug this code:

[paste code]

Error message: [paste error] Expected behavior: [describe] Actual behavior: [describe]

### Code Explanation

Explain this code in simple terms:

• What does it do?
• How does it work step by step?
• What are the key concepts used?

[paste code]

## Analysis Prompts {.cols-2}

### Data Analysis

Analyze this data: [paste data or describe dataset]

Provide:

1. Summary statistics
2. Key patterns and trends
3. Notable outliers
4. Actionable insights
5. Visualization suggestions

### Document Analysis

Analyze this document and provide:

• Executive summary (3-5 sentences)
• Key findings
• Important dates/numbers
• Action items
• Questions that need clarification

[paste document or upload]

### Comparative Analysis

Compare [Option A] vs [Option B]:

• Key similarities
• Key differences
• Pros and cons of each
• Recommendation based on [criteria]

## Image Prompts

### Image Understanding

Analyze this image:

• Describe what you see in detail
• Identify any text
• Note colors, composition, style
• Suggest potential uses

### Image Generation (Imagen)

Generate an image of: [detailed description]

Style: [photorealistic/illustration/cartoon] Mood: [bright/dark/professional] Colors: [color palette] Composition: [layout description]

### Visual Content Ideas

Based on this image, suggest:

• 5 social media caption ideas
• Related content themes
• Color palettes to match
• Similar visual styles to explore

## Productivity Prompts {.cols-2}

### Meeting Preparation

Help me prepare for a meeting about [topic]:

• Key talking points
• Potential questions to expect
• Data points to reference
• Agenda suggestion
• Follow-up action items template

### Research Assistant

Research [topic] and provide:

• Overview (2-3 paragraphs)
• Key facts and statistics
• Recent developments
• Different perspectives
• Reliable sources to explore further

### Learning Helper

Teach me about [topic]:

• Explain core concepts simply
• Use real-world analogies
• Provide practical examples
• Common misconceptions
• Practice questions to test understanding

### Project Planning

Create a project plan for [project]:

• Project phases and milestones
• Task breakdown
• Time estimates
• Resource requirements
• Risk factors and mitigation

## API Usage

### Python SDK

```python
import google.generativeai as genai

genai.configure(api_key="YOUR_API_KEY")
model = genai.GenerativeModel('gemini-1.5-pro')

response = model.generate_content("Hello!")
print(response.text)

With System Instruction

model = genai.GenerativeModel(
    'gemini-1.5-pro',
    system_instruction="You are a helpful coding assistant."
)

response = model.generate_content("Review my code")

Multi-Turn Chat

chat = model.start_chat(history=[])
response = chat.send_message("Hi, I need help with Python")
print(response.text)

response = chat.send_message("How do I read a file?")
print(response.text)

Image Input

import PIL.Image

image = PIL.Image.open('image.jpg')
response = model.generate_content([
    "Describe this image in detail",
    image
])
print(response.text)

Streaming

response = model.generate_content(
    "Write a story",
    stream=True
)
for chunk in response:
    print(chunk.text, end="")

Gemini Advanced Features

Long Context

Process up to 1 million tokens

Analyze this entire codebase/document:
[paste large content]

Identify:
- Overall structure
- Key components
- Dependencies
- Potential issues

Multimodal Input

Combine text, images, and documents

Look at this [image/document] and:
1. Summarize the content
2. Extract key information
3. Answer specific questions about it
4. Suggest improvements

Code Execution

Run Python code in responses

Calculate and show the result:
- Process this data: [data]
- Create a visualization
- Export the results

Keyboard Shortcuts

Web Interface

Best Practices

Do's

• Be specific and detailed
• Provide context and examples
• Use structured formatting
• Break complex tasks into steps
• Iterate and refine prompts
• Verify important information

Don'ts

• Share sensitive/personal data
• Rely solely on AI for critical decisions
• Use vague or ambiguous language
• Expect perfect accuracy
• Skip verification of facts

Tips for Better Results

1. Start with clear objectives
2. Provide relevant context
3. Specify desired format
4. Include constraints
5. Ask follow-up questions
6. Request explanations when needed