Gemini Sheet Orchestrator

Parallel AI prompting system using Google Sheets + Gemini API.

Quick Setup (5 minutes)

1. Get Gemini API Key

• Go to Google AI Studio
• Click "Create API Key"
• Copy the key

2. Create Google Sheet

• Go to Google Sheets
• Name it "Gemini Orchestrator" (or whatever you want)

3. Add the Script

• Click Extensions > Apps Script
• Delete any existing code
• Copy/paste the contents of Code.gs
• Click Save (Ctrl+S)

4. Add API Key

• In Apps Script, click Project Settings (gear icon)
• Scroll to Script Properties
• Click Add Script Property
  • Property: GEMINI_API_KEY
  • Value: (paste your API key)
• Click Save

5. Initialize

• Go back to your spreadsheet
• Refresh the page
• Click Gemini Orchestrator > Setup Sheet
• Authorize when prompted

Usage

Add Tasks

Simply type prompts in column B (Prompt). The system auto-assigns IDs and sets status to "pending".

Run Tasks

• Run All Pending: Processes all pending tasks in parallel batches
• Run Selected Rows: Only runs highlighted rows

Status Values

| Status | Meaning | |--------|---------| | pending | Ready to run | | running | Currently processing | | done | Completed successfully | | error | Failed (check Error column) | | cancelled | Manually cancelled |

Configuration

Edit the Config sheet to adjust:

• Model: gemini-2.0-flash (fast) or gemini-1.5-pro (complex)
• Max Parallel: How many requests run simultaneously (default: 5)
• Temperature: Creativity level 0-1 (default: 0.7)
• Max Tokens: Response length limit (default: 2048)

Auto-Run Mode

To automatically process new tasks:

1. In Apps Script, run setupAutoRun()
2. Tasks will process every 5 minutes automatically

Advanced: Task Dependencies

Add a 10th column called "Depends On" with comma-separated task IDs:

| ID | Prompt | Depends On | |----|--------|------------| | abc123 | Design the database schema | | | def456 | Write the API endpoints | abc123 | | ghi789 | Create the frontend | def456 |

Then use Gemini Orchestrator > Run With Dependencies to process in order.

Example Workflows

Code Generation Pipeline

Row 1: "Design a REST API for a todo app"
Row 2: "Write Python Flask routes for: [output from row 1]"
Row 3: "Create unit tests for: [output from row 2]"
Row 4: "Write documentation for: [output from row 2]"

Content Creation

Row 1: "Generate 5 blog post ideas about AI"
Row 2: "Write outline for: [idea 1]"
Row 3: "Write outline for: [idea 2]"
Row 4: "Write full draft from: [outline 1]"

Research & Analysis

Row 1: "Summarize key trends in renewable energy 2024"
Row 2: "Compare solar vs wind energy costs"
Row 3: "List top 10 companies in each sector"
Row 4: "Create investment thesis from above"

Troubleshooting

"GEMINI_API_KEY not set"

• Check Project Settings > Script Properties

"Quota exceeded"

• Free tier: 60 requests/minute
• Reduce Max Parallel or add delays

"No response generated"

• Check if prompt triggered safety filters
• Try rephrasing the prompt

Cost

• Gemini Flash: Free tier generous, then ~$0.075/1M tokens
• Gemini Pro: ~$1.25/1M input, $5/1M output

For most projects, you'll stay well within free tier.