content-generation

Content Generation

Overview

Generate professional content including images, diagrams, documents, and summaries using AI-powered tools.

Quick Decision Tree

What do you need?
│
├── Generate images
│   └── references/images.md
│   └── Script: scripts/generate_images.py
│
├── Create flowchart/diagram
│   └── references/flowchart.md
│   └── Script: scripts/generate_flowchart.py
│
├── Create Google Doc
│   └── references/document.md
│   └── Script: scripts/generate_document.py
│
├── Generate client proposal
│   └── references/proposal.md
│   └── Script: scripts/generate_proposal.py
│
├── Convert Markdown to PDF
│   └── references/pdf.md
│   └── Script: scripts/md_to_pdf.py
│
└── Summarize content
    └── references/summarize.md
    └── Script: scripts/summarize_content.py

Environment Setup

# Required in .env
FAL_API_KEY=fal_xxxx              # For image generation
OPENROUTER_API_KEY=sk-or-v1-xxx   # For AI text generation

Get API keys:

• fal.ai: https://fal.ai/dashboard/keys
• OpenRouter: https://openrouter.ai/keys

Common Usage

Generate Images

# Nano Banana Pro (best quality)
python scripts/generate_images.py "Professional product photo of smartwatch" --model nano-banana-pro

# FLUX-2 (fast, cheap)
python scripts/generate_images.py "Vibrant lifestyle photo" --model flux-2 --size landscape_4_3

Create Flowchart

python scripts/generate_flowchart.py "User login: enter email, validate, check password, if correct go to dashboard"

Generate Document

python scripts/generate_document.py --input content.json --title "Q4 Report"

Generate Proposal

python scripts/generate_proposal.py --transcript-file meeting.txt --client "Acme Corp"

Cost Estimates

Tool	Cost
Nano Banana Pro 1K	$0.15/image
Nano Banana Pro 4K	$0.30/image
FLUX-2	$0.012/megapixel
Flowchart	~$0.005/diagram
Proposal	~$0.10/doc

Security Notes

Credential Handling

• Store FAL_API_KEY in .env file (never commit to git)
• Store OPENROUTER_API_KEY in .env file (never commit to git)
• Regenerate keys from respective dashboards if compromised
• Never log or print API keys in script output

Data Privacy

• Text prompts sent to fal.ai and OpenRouter servers
• Generated images stored locally in .tmp/ directory
• AI services may log prompts for service improvement
• Avoid including sensitive/confidential data in prompts
• Generated documents may be uploaded to Google Drive

Access Scopes

• FAL_API_KEY - Full access to image generation models
• OPENROUTER_API_KEY - Access to configured AI models
• Google OAuth - Required for Google Docs creation (see google-workspace skill)

Compliance Considerations

• AI Disclosure: Consider disclosing AI-generated content where required
• Image Rights: Generated images subject to fal.ai terms of use
• Copyright: AI-generated content copyright varies by jurisdiction
• Brand Consistency: Review AI outputs against brand guidelines
• Sensitive Content: AI image generators have content policies
• Client Proposals: Verify accuracy before sending to clients
• Attribution: Some contexts require AI generation disclosure

Troubleshooting

Common Issues

Issue: Image generation failed

Symptoms: fal.ai returns error or no image generated Cause: Invalid prompt, model unavailable, or API issue Solution:

• Check prompt for policy violations (explicit content, etc.)
• Verify model ID is correct and available
• Try a simpler prompt to test API connectivity
• Check fal.ai status page for outages

Issue: Content policy violation

Symptoms: "Content policy violation" or similar safety error Cause: Prompt contains flagged terms or concepts Solution:

• Remove explicit, violent, or harmful content from prompt
• Rephrase ambiguous terms that might trigger filters
• Review fal.ai content policy guidelines
• Use more general/professional language

Issue: Google Doc not created

Symptoms: Script completes but no document appears in Drive Cause: OAuth issue, folder permissions, or API error Solution:

• Verify Google OAuth is working (see google-workspace skill)
• Check target folder exists and is accessible
• Delete mycreds.txt and re-authenticate
• Look for error messages in script output

Issue: Flowchart syntax error

Symptoms: Mermaid diagram fails to render Cause: Invalid Mermaid syntax in generated diagram Solution:

• Review generated Mermaid code for syntax errors
• Simplify complex flows into smaller diagrams
• Test diagram at https://mermaid.live/
• Ensure special characters are properly escaped

Issue: PDF conversion fails

Symptoms: Error during markdown to PDF conversion Cause: Missing dependencies or invalid markdown Solution:

• Ensure required PDF tools are installed (weasyprint, etc.)
• Check markdown syntax is valid
• Verify font files exist if custom fonts specified
• Try with simpler markdown to isolate issue

Issue: OpenRouter API error

Symptoms: Text generation fails with API error Cause: Invalid API key, model unavailable, or rate limit Solution:

• Verify OPENROUTER_API_KEY is set correctly
• Check model availability at OpenRouter dashboard
• Try a different model if current one is unavailable
• Review rate limit status and add delays if needed

Resources

• references/images.md - fal.ai image generation
• references/flowchart.md - Mermaid diagram generation
• references/document.md - Google Docs creation
• references/proposal.md - Client proposal generation
• references/pdf.md - Markdown to PDF conversion
• references/summarize.md - Content summarization

Integration Patterns

Generate and Save

Skills: content-generation → google-workspace Use case: Create documents and store in Drive Flow:

1. Generate content (proposal, report, summary)
2. Format as Google Doc or PDF
3. Upload to appropriate Drive folder via google-workspace

Proposal to CRM

Skills: content-generation → attio-crm Use case: Link generated proposals to deal records Flow:

1. Generate client proposal from meeting transcript
2. Upload proposal to Google Drive
3. Add note to Attio CRM company record with proposal link

Images to Video

Skills: content-generation → video-production Use case: Create title slides for video courses Flow:

1. Generate branded title slide images for each lesson
2. Export images to video-compatible format
3. Use video-production to stitch with lesson videos