Multimodal Prompting

This prompt provides a clear framework for image analysis. The model knows exactly what information you need.

Analyze this image and describe:

1. **Main Subject**: What is the primary focus of this image?
2. **Setting**: Where does this appear to be? (indoor/outdoor, location type)
3. **Mood**: What emotional tone or atmosphere does it convey?
4. **Text Content**: Any visible text, signs, or labels?
5. **Notable Details**: What might someone miss at first glance?
6. **Technical Quality**: How is the lighting, focus, and composition?

[Paste or describe the image you want to analyze]

Image description or URL: ${imageDescription}

Get structured data from image analysis that's easy to parse and use in applications.

Analyze this image and return a JSON object with the following structure:

{
"summary": "One sentence description",
"objects": ["List of main objects visible"],
"people": {
  "count": "number or 'none'",
  "activities": ["What they're doing, if any"]
},
"text_detected": ["Any text visible in the image"],
"colors": {
  "dominant": ["Top 3 colors"],
  "mood": "Warm/Cool/Neutral"
},
"setting": {
  "type": "indoor/outdoor/unknown",
  "description": "More specific location description"
},
"technical": {
  "quality": "high/medium/low",
  "lighting": "Description of lighting",
  "composition": "Description of framing/composition"
},
"confidence": "high/medium/low"
}

Image to analyze: ${imageDescription}

Compare two or more images with specific criteria that matter to your decision.

Compare these images for ${purpose}:

**Image A**: ${imageA}
**Image B**: ${imageB}

Analyze each image on these criteria:
1. ${criterion1} (importance: high)
2. ${criterion2} (importance: medium)  
3. ${criterion3} (importance: low)

Provide:
- Side-by-side comparison for each criterion
- Strengths and weaknesses of each
- Clear recommendation with reasoning
- Any concerns or caveats

Extract structured data from photos of documents, receipts, invoices, or forms.

This is a photo/scan of a ${documentType}.

Extract all information into structured JSON format:

{
"document_type": "detected type",
"date": "if present",
"key_fields": {
  "field_name": "value"
},
"line_items": [
  {"description": "", "amount": ""}
],
"totals": {
  "subtotal": "",
  "tax": "",
  "total": ""
},
"handwritten_notes": ["any handwritten text"],
"unclear_sections": ["areas that were hard to read"],
"confidence": "high/medium/low"
}

IMPORTANT: If any text is unclear, note it in "unclear_sections" rather than guessing. Mark confidence as "low" if significant portions were hard to read.

Document description: ${documentDescription}

Get detailed analysis of screenshots for debugging, UX review, or documentation.

This is a screenshot of ${applicationName}.

Analyze this interface:

**Identification**
- What screen/page/state is this?
- What is the user likely trying to accomplish here?

**UI Elements**
- Key interactive elements (buttons, forms, menus)
- Current state (anything selected, filled in, or expanded?)
- Any error messages, warnings, or notifications?

**UX Assessment**
- Is the layout clear and intuitive?
- Any confusing elements or unclear labels?
- Accessibility concerns (contrast, text size, etc.)?

**Issues Detected**
- Visual bugs or misalignments?
- Truncated text or overflow issues?
- Inconsistent styling?

Screenshot description: ${screenshotDescription}

Get plain-language explanations and fixes for error messages in screenshots.

I'm seeing this error in ${context}.

[Describe or paste the error message/screenshot]
Error details: ${errorDetails}

Please provide:

1. **Plain Language Explanation**: What does this error actually mean?

2. **Likely Causes** (ranked by probability):
 - Most likely: 
 - Also possible:
 - Less common:

3. **Step-by-Step Fix**:
 - First, try...
 - If that doesn't work...
 - As a last resort...

4. **Prevention**: How to avoid this error in the future

5. **Red Flags**: When this error might indicate a more serious problem

Use this template to create detailed, specific image generation prompts.

Create an image with these specifications:

**Subject**: ${subject}

**Style**: ${style}
**Medium**: ${medium} (e.g., oil painting, digital art, photograph)

**Composition**:
- Framing: ${framing} (close-up, medium shot, wide angle)
- Perspective: ${perspective} (eye level, low angle, overhead)
- Focus: ${focusArea}

**Lighting**:
- Source: ${lightSource}
- Quality: ${lightQuality} (soft, harsh, diffused)
- Time of day: ${timeOfDay}

**Color Palette**: ${colors}

**Mood/Atmosphere**: ${mood}

**Must Include**: ${includeElements}
**Must Avoid**: ${avoidElements}

**Technical**: ${aspectRatio} aspect ratio, high quality

Build complex scenes by describing what appears in each layer of depth.

Generate a detailed scene:

**Setting**: ${setting}

**Foreground** (closest to viewer):
${foreground}

**Middle Ground** (main action area):
${middleGround}

**Background** (distant elements):
${background}

**Atmospheric Details**:
- Weather/Air: ${weather}
- Lighting: ${lighting}
- Time: ${timeOfDay}

**Style**: ${artisticStyle}
**Mood**: ${mood}
**Color Palette**: ${colors}

Additional details to include: ${additionalDetails}

Get accurate transcriptions with speaker labels, timestamps, and handling of unclear sections.

Transcribe this audio recording.

**Context**: ${recordingType} (meeting, interview, podcast, lecture, etc.)
**Expected Speakers**: ${speakerCount} (${speakerRoles})
**Domain**: ${domain} (technical terms to expect: ${technicalTerms})

**Output Format**:
[00:00] **Speaker 1 (Name/Role)**: Transcribed text here.
[00:15] **Speaker 2 (Name/Role)**: Their response here.

**Instructions**:
- Include timestamps at natural breaks (every 30-60 seconds or at speaker changes)
- Mark unclear sections as [inaudible] or [unclear: best guess?]
- Note non-speech sounds in brackets: [laughter], [phone ringing], [long pause]
- Preserve filler words only if they're meaningful (um, uh can be removed)
- Flag any action items or decisions with → symbol

Audio description: ${audioDescription}

Get a comprehensive analysis of audio content including summary, key moments, and sentiment.

Analyze this audio recording:

Audio description: ${audioDescription}

Provide:

**1. Executive Summary** (2-3 sentences)
What is this recording about? What's the main takeaway?

**2. Speakers**
- How many distinct speakers?
- Characteristics (if discernible): tone, speaking style, expertise level

**3. Content Breakdown**
- Main topics discussed (with approximate timestamps)
- Key points made
- Questions raised

**4. Emotional Analysis**
- Overall tone (formal, casual, tense, friendly)
- Notable emotional moments
- Energy level throughout

**5. Actionable Items**
- Decisions made
- Action items mentioned
- Follow-ups needed

**6. Notable Quotes**
Pull out 2-3 significant quotes with timestamps

**7. Audio Quality**
- Overall clarity
- Any issues (background noise, interruptions, technical problems)

Get a structured breakdown of video content including timeline, visual elements, and key moments.

Analyze this video: ${videoDescription}

Provide a comprehensive analysis:

**1. Overview** (2-3 sentences)
What is this video about? What's the main message or purpose?

**2. Timeline of Key Moments**
| Timestamp | Event | Significance |
|-----------|-------|--------------|
| 0:00 | ... | ... |

**3. Visual Analysis**
- Setting/Location: Where does this take place?
- People: Who appears? What are they doing?
- Objects: Key items or props featured
- Visual style: Quality, editing, graphics used

**4. Audio Analysis**
- Speech: Main points made (if any dialogue)
- Music: Type, mood, how it's used
- Sound effects: Notable audio elements

**5. Production Quality**
- Video quality and editing
- Pacing and structure
- Effectiveness for its purpose

**6. Target Audience**
Who is this video made for? Does it serve them well?

**7. Key Takeaways**
What should a viewer remember from this video?

Extract specific information from videos with timestamps and structured output.

Extract specific information from this video:

Video type: ${videoType}
Video description: ${videoDescription}

**Information to Extract**:
1. ${extractItem1}
2. ${extractItem2}
3. ${extractItem3}

**Output Format**:
{
"video_summary": "Brief description",
"duration": "estimated length",
"extracted_data": [
  {
    "timestamp": "MM:SS",
    "item": "What was found",
    "details": "Additional context",
    "confidence": "high/medium/low"
  }
],
"items_not_found": ["List anything requested but not present"],
"additional_observations": "Anything relevant not explicitly requested"
}

Verify that images accurately represent their text descriptions and vice versa.

Analyze this image and its accompanying text for alignment:

**Image**: ${imageDescription}
**Text Description**: "${textDescription}"

Evaluate:

**1. Accuracy Match**
- Does the image show what the text describes?
- Score: [1-10] with explanation

**2. Text Claims vs. Visual Reality**
| Claim in Text | Visible in Image? | Notes |
|---------------|-------------------|-------|
| ... | Yes/No/Partial | ... |

**3. Visual Elements Not Mentioned**
What's visible in the image but not described in the text?

**4. Text Claims Not Visible**
What's described in text but can't be verified from the image?

**5. Recommendations**
- For the text: [improvements to match image]
- For the image: [improvements to match text]

**6. Overall Assessment**
Is this image-text pair trustworthy for ${purpose}?

Debug UI issues by analyzing both the visual output and the source code together.

I have a UI bug. Here's what I see and my code:

**Screenshot Description**: ${screenshotDescription}
**What's Wrong**: ${bugDescription}
**Expected Behavior**: ${expectedBehavior}

**Relevant Code**:
```${language}
${code}
```

Please help me:

**1. Root Cause Analysis**
- What in the code is causing this visual issue?
- Which specific line(s) are responsible?

**2. Explanation**
- Why does this code produce this visual result?
- What's the underlying mechanism?

**3. The Fix**
```${language}
// Corrected code here
```

**4. Prevention**
- How to avoid this type of bug in the future
- Any related issues to check for

Compare multiple images systematically against your criteria to make informed decisions.

I'm choosing between these options for ${purpose}:

**Option A**: ${optionA}
**Option B**: ${optionB}
**Option C**: ${optionC}

**My Criteria** (in order of importance):
1. ${criterion1} (weight: high)
2. ${criterion2} (weight: medium)
3. ${criterion3} (weight: low)

Provide:

**Comparison Matrix**
| Criterion | Option A | Option B | Option C |
|-----------|----------|----------|----------|
| ${criterion1} | Score + notes | ... | ... |
| ${criterion2} | ... | ... | ... |
| ${criterion3} | ... | ... | ... |

**Weighted Scores**
- Option A: X/10
- Option B: X/10
- Option C: X/10

**Recommendation**
Based on your stated priorities, I recommend [Option] because...

**Caveats**
- If [condition], consider [alternative] instead
- Watch out for [potential issue]

This prompt explicitly handles cases where the model can't see clearly or is uncertain.

Analyze this image: ${imageDescription}

**Instructions for Handling Uncertainty**:

IF YOU CAN'T SEE SOMETHING CLEARLY:
- Don't guess or make up details
- Say: "I can see [what's visible] but cannot clearly make out [unclear element]"
- Suggest what additional information would help

IF CONTENT SEEMS RESTRICTED:
- Explain what you can and cannot analyze
- Focus on permitted aspects of the analysis

IF ASKED ABOUT PEOPLE:
- Describe actions, positions, and general characteristics
- Do not attempt to identify specific individuals
- Focus on: number of people, activities, expressions, attire

**Your Analysis**:
[Proceed with analysis, applying these guidelines]