How fresh is the data the AI sees?

By default, chat reads see data at most 60 seconds old (TTL gate); report widgets at most 5 minutes. Drive change notifications bypass the TTL — typical workbook-edit-to-mirror latency is 5-10 seconds. You can also set the source to 'always live' for sub-second.

What happens to formulas?

Sheets returns calculated values by default — that's what we mirror. The cell's underlying formula isn't stored; only its result. If a formula references an external API (`IMPORTRANGE`, `IMPORTDATA`), the value can lag the upstream source by ~5 minutes (Sheets' own caching).

Can I edit my Sheet from iDBQuery?

Read-only by default. Optional opt-in write-back lands in Wave D (append rows / update cells / create tab). You'll need to enable it per-source via the source detail page; every write is audited.

How big a workbook can I connect?

Hard caps: 50 tabs per source, 1M rows per tab. Practical: a 500K-row workbook syncs in ~10 seconds and the SQLite mirror weighs ~50 MB. Larger needs split across multiple sources.

Spreadsheet

Google Sheets

Your spreadsheet, but it actually answers questions — live.

Connect free Talk to us

Overview

Why teams connect Google Sheets to iDBQuery

iDBQuery's Google Sheets connector mirrors selected tabs of a workbook into a per-source SQLite cache, then keeps it in sync via TTL-gated reads (60-second default for chat, 5-minute for reports) plus Drive change-notification webhooks. Edits in the workbook show up in your dashboard within seconds, no re-upload required. The connector requests `drive.file` + `spreadsheets.readonly` — only files you explicitly grant. Cross-source DuckDB federation lets you join your Sheets data against MySQL, PostgreSQL, BigQuery, or any other connected source in one chat.

Capabilities

What you can do

Everything iDBQuery's Google Sheets connector supports out of the box.

OAuth 2.0 with `drive.file` scope — no broad Drive access
Per-tab selection with row-count + last-modified signals
Type inference per column (INTEGER / REAL / TEXT)
Hash-diff sync — only changed rows re-written
TTL freshness gate (60s chat / 300s report defaults; 'always live' override)
Drive webhook bypass — edits land in mirror within 5-10s
Cross-source federation via DuckDB (Sheets × Postgres × MySQL)
AI tools: sheets_describe_workbook, sheets_refresh_now, sheets_change_summary
Schema-memory auto-refresh after every sync

Setup

Get connected in minutes

Pick Google Sheets in the connection picker

From your project, click 'Add source' → Google Sheets. Sign in with Google in the popup. iDBQuery requests only the read scopes (`drive.file` + `spreadsheets.readonly`).

Pick the workbook

Search the workbooks Google has shared with the iDBQuery app. To make a workbook visible, share it with this app via Google Drive's share dialog.

Pick the tabs to sync

Multi-select from the workbook's tabs. Hidden tabs are skipped by default. Optional 'always live' toggle for sub-second TTL.

Open chat

Ask 'show me last week's pipeline by region' — iDBQuery generates SQL against the mirrored tabs and returns a chart. Edits in Sheets reflect on the next read.

Try this

Sample prompts

>What's the largest deal in Q1 Pipeline?
>Sum revenue by month from the Budget tab
>Compare actual to forecast across departments
>Did anything change in the workbook today?
>Refresh now and tell me what's new

Frequently asked questions

No. We use the `drive.file` scope, which means only files you explicitly grant via Google's share dialog (or pick via the OAuth flow) are accessible to iDBQuery. We can never list or read other files in your Drive.

Keep exploring

Chat with data Reports All integrations Finance & ops use case

Connect Google Sheets in under a minute.

Free tier covers most teams. No credit card.

Start free Contact us

No credit card required · 1M tokens / month free

Loading…

Frequently asked questions