Connect AI Agents to Google Sheets: Read, Write, Create via API

AI agents are increasingly used to automate research, data collection, and reporting workflows. One of the most common needs: storing and sharing structured data. Google Sheets is the go-to collaborative spreadsheet — and now IteraTools lets any AI agent create, read, and write spreadsheets via a single API call.

In this tutorial, we'll walk through a complete agent workflow:

1. Create a spreadsheet with headers
2. Write research data to it
3. Read it back as structured JSON
4. Share the URL with a human

The API

IteraTools is a consumption-based API for AI agents — you pay per call with no monthly fee. Three new endpoints handle Google Sheets:

Endpoint	Price	What it does
POST /sheets/create	$0.005	Create a new spreadsheet, get URL
POST /sheets/write	$0.003	Write or append rows
GET /sheets/read	$0.002	Read rows as JSON objects

Step 1: Create a spreadsheet

curl -X POST https://api.iteratools.com/sheets/create \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "title": "Market Research Q1 2025",
    "headers": ["Company", "Location", "Revenue", "Notes"]
  }'

Response:

{
  "ok": true,
  "data": {
    "spreadsheet_id": "1abc...xyz",
    "url": "https://docs.google.com/spreadsheets/d/1abc...xyz/edit",
    "title": "Market Research Q1 2025"
  }
}

The spreadsheet is automatically set to public read-only so you can share the URL immediately. Save the spreadsheet_id for the next steps.

Step 2: Write research data

curl -X POST https://api.iteratools.com/sheets/write \
  -H "Authorization: Bearer YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "spreadsheet_id": "1abc...xyz",
    "values": [
      ["Acme Corp", "São Paulo", "R$ 12M", "Growing fast, contact CEO"],
      ["TechBR", "Florianópolis", "R$ 4M", "SaaS, looking for partnerships"],
      ["DataSul", "Curitiba", "R$ 8M", "Enterprise focus, good fit"]
    ],
    "mode": "append"
  }'

Response:

{
  "ok": true,
  "data": {
    "updated_rows": 3
  }
}

Use mode: "append" to add rows below existing data, or mode: "overwrite" (default) to replace from the starting cell.

Step 3: Read the data back

curl "https://api.iteratools.com/sheets/read?spreadsheet_id=1abc...xyz" \
  -H "Authorization: Bearer YOUR_KEY"

Response:

{
  "ok": true,
  "data": {
    "headers": ["Company", "Location", "Revenue", "Notes"],
    "rows": [
      {"Company": "Acme Corp", "Location": "São Paulo", "Revenue": "R$ 12M", "Notes": "Growing fast, contact CEO"},
      {"Company": "TechBR", "Location": "Florianópolis", "Revenue": "R$ 4M", "Notes": "SaaS, looking for partnerships"},
      {"Company": "DataSul", "Location": "Curitiba", "Revenue": "R$ 8M", "Notes": "Enterprise focus, good fit"}
    ],
    "total": 3,
    "range": "Sheet1!A1:D4"
  }
}

The first row is automatically treated as headers, and all subsequent rows are returned as key-value objects — ready to use in your agent's logic without any parsing.

Full Python agent example

Here's a complete example of an agent that does market research and stores results in a spreadsheet:

import requests

API_KEY = "your_iteratools_key"
BASE = "https://api.iteratools.com"
HEADERS = {"Authorization": f"Bearer {API_KEY}", "Content-Type": "application/json"}

# 1. Create a spreadsheet
resp = requests.post(f"{BASE}/sheets/create", headers=HEADERS, json={
    "title": "Agent Research Results",
    "headers": ["Company", "Website", "Category", "Score"]
})
sheet_id = resp.json()["data"]["spreadsheet_id"]
sheet_url = resp.json()["data"]["url"]
print(f"📊 Spreadsheet created: {sheet_url}")

# 2. Simulate agent gathering data (in real use, this would call /search, /scrape, etc.)
research_results = [
    ["Startup A", "startupa.com", "SaaS", 8.5],
    ["Startup B", "startupb.io", "Marketplace", 7.2],
    ["Startup C", "startupc.com.br", "Fintech", 9.1],
]

# 3. Write to spreadsheet
resp = requests.post(f"{BASE}/sheets/write", headers=HEADERS, json={
    "spreadsheet_id": sheet_id,
    "values": research_results,
    "mode": "append"
})
print(f"✅ Wrote {resp.json()['data']['updated_rows']} rows")

# 4. Read back to verify
resp = requests.get(f"{BASE}/sheets/read", headers=HEADERS, params={"spreadsheet_id": sheet_id})
data = resp.json()["data"]
print(f"\n📋 Research results ({data['total']} companies):")
for row in data["rows"]:
    print(f"  • {row['Company']} ({row['Category']}) — Score: {row['Score']}")

print(f"\n🔗 Share link: {sheet_url}")

Output:

📊 Spreadsheet created: https://docs.google.com/spreadsheets/d/1abc...xyz/edit
✅ Wrote 3 rows

📋 Research results (3 companies):
  • Startup A (SaaS) — Score: 8.5
  • Startup B (Marketplace) — Score: 7.2
  • Startup C (Fintech) — Score: 9.1

🔗 Share link: https://docs.google.com/spreadsheets/d/1abc...xyz/edit

Why this matters for agents

The typical agent loop is: search → scrape → process → store → report. The "store and report" step used to require setting up a database, a reporting tool, or emailing CSV files. With /sheets/create, an agent can instantly produce a shareable, human-readable spreadsheet — no infrastructure needed.

Combined with IteraTools' other endpoints:

• /search to find companies
• /scrape to extract data from their websites
• /sheets/create + /sheets/write to store results
• /tts or /email/send to notify stakeholders

You have a complete research pipeline for under $0.10 total.

Getting started

1. Get an API key at iteratools.com
2. Call POST /sheets/create to create your first spreadsheet
3. Use the returned spreadsheet_id for all subsequent reads and writes

Full documentation: iteratools.com/docs

---

IteraTools is a consumption-based API with 41 tools designed for AI agents — image generation, web scraping, TTS, WhatsApp, PDF processing, and now Google Sheets. No subscriptions, pay per call.