Best Gemini Prompts

Analyze the attached {{image_type}} and provide a comprehensive breakdown:

1. **Visual inventory**: List every distinct element you can identify (objects, text, colors, layout)
2. **Text extraction**: Transcribe ALL text visible in the image exactly as written
3. **Spatial relationships**: Describe how elements are positioned relative to each other
4. **Context clues**: What can you infer about when, where, and why this was created?
5. **Data extraction**: If this contains charts, tables, or diagrams, extract the data into a structured {{output_format}} format

For any element you're uncertain about, say "[uncertain]" rather than guessing.

Finally, suggest 3 follow-up questions I could ask about this image to get deeper insights.

I'm providing a {{document_type}} that is approximately {{page_count}} pages long.

Summarize it at three levels of detail:

**Level 1 — Executive Summary (3 sentences max)**
The single most important takeaway, who should care, and what action to take.

**Level 2 — Section-by-Section (one paragraph each)**
For each major section or chapter, provide: key argument, supporting evidence, and how it connects to the overall thesis.

**Level 3 — Detailed Notes**
Bullet points of every specific claim, data point, quote, or recommendation worth remembering. Tag each bullet as [Fact], [Claim], [Quote], or [Recommendation].

Also flag:
- Any contradictions within the document
- Claims that lack supporting evidence
- The strongest and weakest arguments

Maintain the document's original terminology — do not paraphrase technical terms.

Generate a {{language}} implementation for {{feature_description}}.

Project context:
- Framework: {{framework}}
- Existing patterns: {{pattern_description}}
- This code will be used in: {{usage_context}}

Requirements:
1. Follow {{framework}} best practices and idioms
2. Include comprehensive error handling with typed errors
3. Add TypeScript types / type hints for all parameters and return values
4. Write the code in a way that's testable (dependency injection where appropriate)
5. Include inline comments ONLY for non-obvious logic

Output format:
- Main implementation file
- Types/interfaces file (if needed)
- One example usage showing how to call this code

Do not include package installation instructions or boilerplate I didn't ask for.

Parse the following {{input_type}} and return a structured JSON response.

Input:
{{input_data}}

Output schema:
{
  "entities": [
    {
      "name": "string",
      "type": "{{entity_type}}",
      "attributes": { },
      "relationships": [{ "target": "string", "relation": "string" }],
      "confidence": "high | medium | low"
    }
  ],
  "summary": "string",
  "metadata": {
    "input_quality": "clean | noisy | ambiguous",
    "entities_found": "number",
    "processing_notes": "string"
  }
}

Rules:
- Return ONLY valid JSON, no markdown code fences, no explanation text
- If a field cannot be determined, use null (not empty string)
- Set confidence to "low" for any entity extracted from ambiguous context
- Deduplicate entities that appear multiple times with different names

I need help with a Google {{workspace_tool}} task.

Current state: {{current_state}}
Goal: {{desired_outcome}}

Provide:
1. **Step-by-step instructions** using the actual Google {{workspace_tool}} UI (reference menu paths like File > Download > CSV)
2. **Formulas/scripts** if applicable — use Google Apps Script (not Excel VBA) syntax
3. **Automation opportunity**: Can this be automated with Apps Script? If yes, provide a complete script with:
   - Trigger setup instructions (time-based, on-edit, etc.)
   - Error handling and logging
   - A description of what to modify for different use cases
4. **Template**: If this is a repeating task, create a template structure I can reuse

Format any formulas as code blocks. For Apps Script, include the full function — not snippets.

Extract structured data from the attached {{document_type}}.

Extract these fields:
{{fields_list}}

Output as a {{output_format}} with one row per {{entity_unit}}.

Extraction rules:
- If a field appears multiple times, take the most recent / most specific value
- If a field is missing, use "NOT_FOUND" (not null, not empty)
- For dates, normalize to YYYY-MM-DD format regardless of input format
- For currency, normalize to numbers without symbols (include a "currency" column)
- For names, use "LastName, FirstName" format
- Flag any field where the extraction is ambiguous with a trailing " [AMBIGUOUS]" marker

After the data, provide:
- Total records extracted
- Fields with the highest "NOT_FOUND" rate
- Any patterns or anomalies you noticed in the data

Analyze the attached {{media_type}} and provide:

1. **Timeline summary**: Key moments with timestamps (format: [MM:SS])
   - What happens at each moment
   - Any text or graphics that appear on screen
   - Speaker changes (if multiple speakers)

2. **Full transcript** (if audio/spoken content exists):
   - Include speaker labels: [Speaker 1], [Speaker 2], etc.
   - Note [inaudible] segments and [background noise/music]
   - Include timestamps every 30 seconds

3. **Content analysis**:
   - Main topics discussed (with timestamp ranges)
   - Key claims or statements made
   - Action items or decisions mentioned
   - Tone/sentiment shifts throughout

4. **Searchable index**: Create a topic index so I can jump to specific parts:
   Topic -> Timestamp -> One-line summary

Research {{topic}} using current information.

I need:
1. **Current facts**: What is the latest information as of today? Include specific numbers, dates, and sources.
2. **Recent developments**: What changed in the last {{timeframe}}? List chronologically.
3. **Key players**: Who are the main companies/people involved and what's their current position?
4. **Consensus vs debate**: What do most sources agree on? Where is there disagreement?
5. **Primary sources**: For each major claim, identify the original source (not a summary article)

Quality requirements:
- Distinguish between confirmed facts and reported/rumored information
- Note if a source has a potential bias (e.g., company blog vs independent research)
- If information is rapidly changing, note the date of the most recent data point
- Do NOT include information you're unsure about — gaps are better than errors

End with 3 specific search queries I can use to verify your most important claims.