The AI engine for Google Sheets™ & Docs™

Never see “Loading…”
ever again.

SheetEngine runs AI on thousands of rows as server-side jobs — the exact cost shown before you click Run, results that land even after you close the tab, and zero token markup on your own keys.

PRIVATE BETA · SUMMER 2026 · first 100 seats lock founder pricing at $99/yr

0 timeoutsJobs run on our servers, not in fragile formulas
$0.00 markupBring your own key, pay flat — we never touch your tokens
0k+ rowsPer job, resumable, with per-row retry
1-click=GPT() migration — your old sheets work day one

Documented failure modes of formula-based AI add-ons

Cells stuck on “Loading…” for hours Formula re-ran and charged me twice 30-second Apps Script timeout killed the run No idea what the job costs until it's done Credits vanished from the usage screen Recalculation burned my whole credit pack Closed the tab — lost the whole run $800/month for enrichment credits Docs output is a wall of unformatted text No scheduled runs, no triggers, no API

Radical cost transparency

You'll know the price before you ever click Run.

Every bulk job starts with a dry run: five sample rows processed for real, actual outputs shown, and the projected total for the full range — down to the cent.

  • Live receipts after every job — rows, cost, duration, failures: zero.
  • Results written as values. Recalculation can never re-bill you.
  • Spend caps per user, per team, per month. No surprises, ever.
SHEETENGINE · DRY RUN
Enrich: leads_q3.gsheet claude-haiku
acme-metals.com
brightlane.io
nordvik.se
12,400 rows · projected ≈ $3.82

Built different, on purpose

Four reasons teams switch — and never go back.

01 / RELIABILITY

A job engine, not a fragile formula

Bulk work runs on our servers — queued, resumable, idempotent. No 30-second Apps Script timeout. Close the laptop; the results are in your cells when you're back. Failed rows retry automatically and report in a per-job log.

// never pay for the same cell twice
02 / TRANSPARENCY

The price, before — not after

Dry-run preview on every job with real sample outputs and the projected total. Live spend meter, per-job receipts, team caps. Managed keys bill on a published per-model rate card — never opaque credits.

// "12,400 rows ≈ $3.82 — run?"
03 / AUTOMATION

AI that runs without you

Schedule it: "re-translate the catalog every Monday, 6am." Trigger it: new row lands → classify + enrich. Every sheet becomes an API endpoint, with first-class n8n and Zapier integrations at launch.

// a tool you open becomes infrastructure you depend on
04 / DOCS, DONE RIGHT

Mail-merge with a brain

Point a Doc template at a Sheet → generate N polished, formatted Docs or PDFs in Drive — proposals, briefs, outreach one-pagers. Real headings, tables, and styles. Preview every change before it applies.

// rows in, documents out
=AI() =AI_ENRICH() =AI_TRANSLATE() =AI_CLASSIFY() =AI_EXTRACT() =AI_SEARCH() =AI_VISION() =AI_SUMMARIZE() =AI_REWRITE() =GPT() → works day one

every function · every major model — OpenAI, Claude, Gemini, DeepSeek, Mistral, Grok — pick per task

The moat nobody else has

Your spreadsheet keeps working
after you stop.

Schedule

On the clock

Recurring jobs that re-run themselves — catalog translation, enrichment refresh, weekly report prep.

$ every monday 06:00
→ re-translate catalog (8,212 rows)
✓ done before you wake up
Trigger

On new rows

A form response or imported lead lands in the sheet → it's classified, enriched, and scored within seconds.

$ on new_row in leads!A:A
→ =AI_ENRICH(company) → ICP score
✓ row 4,081 enriched · $0.0003
Webhook · n8n · Zapier

On your stack

Every sheet becomes an AI endpoint. Native n8n node and Zapier app from day one — not an afterthought.

$ POST /v1/sheets/:id/jobs
→ { "tool": "classify", "rows": "*" }
✓ 202 queued · webhook on complete

Sheets → Docs bridge

One sheet in.
A hundred documents out.

Each row becomes a personalized, properly formatted proposal, brief, or report in Drive — headings, tables, and brand styles intact. Nobody else owns this. We will.

The honest comparison

Same surface. Different machine.

We publish this because transparency is the product. Judge for yourself.

SheetEngineGPT for SheetsClay
Pricing basisFlat $12/mo BYOK · $0 markupCredit packs $29–$299+Credits, from $149/mo
Cost shown before the run✓ every jobpartial
Survives the 30s formula timeout✓ server-side jobs✗ Apps Script limitn/a — not in Sheets
Jobs continue after tab closes
Recalculation can re-bill younever — values + dedup ledgerdocumented issuen/a
Scheduled runs & new-row triggers✓ built-in
Native in Google Sheets✗ separate app
Sheets → formatted Docs/PDFs✓ merge-generatelimited
=GPT() formula compatibility✓ migration mode✓ (it's theirs)

Based on publicly listed pricing and documentation, June 2026. Names belong to their owners; we just think you deserve the table.

Pricing without the fog

Flat. Published. Boring on purpose.

Free

$0

Kick the tires properly.

  • 300 AI operations / month
  • All functions & bulk tools
  • Cost preview on every job
  • Managed keys only
JOIN WAITLIST
FOUNDER PRICE · FIRST 100

BYOK Pro

$12/user/mo · $99/yr

Your API keys. Our engine. Zero markup, forever.

  • Unlimited ops on your keys
  • $0.00 token markup
  • Schedules, triggers, webhooks
  • =GPT() migration mode
LOCK FOUNDER PRICING

Managed Pro

usage · rate card +20%

We handle the keys. You see the exact math.

  • Published per-model rate card
  • Per-job receipts
  • Monthly spend caps
  • All models, one bill
JOIN WAITLIST

Team

$29/mo base + usage

For pods that run pipelines together.

  • Shared prompt library
  • Pooled credits + member caps
  • Priority job queue
  • Admin spend controls
JOIN WAITLIST

THE MANAGED RATE CARD IS PUBLIC: provider price + 20%, listed per model. NO CREDITS. NO MYSTERY MATH.

Fair questions.

When can I use it?
Private beta opens summer 2026, rolling out to the waitlist in order. The first 100 seats lock founder pricing — $99/year for BYOK Pro, forever.
Do I need my own API key?
Your choice. BYOK: bring an OpenAI, Anthropic, or Google key, pay a flat subscription, and we never mark up a single token. Managed: we handle keys and bill on a published per-model rate card.
Which models are supported?
OpenAI, Claude, Gemini, DeepSeek, Mistral, and Grok at launch — selectable per task, with side-by-side A/B comparison so you can pick the cheapest model that's good enough.
Will my existing =GPT() formulas work?
Yes. Migration mode accepts =GPT() signatures as aliases, so sheets built with GPT for Sheets work day one — plus a one-click importer for your saved prompts.
What happens to my data?
Nothing is stored. Cell content passes through to the model provider and back. We retain only content hashes (for caching and dedup billing) and job metadata. Zero data retention by design — see the privacy policy.
How is this different from Clay?
Clay is a powerful separate app with credit pricing from $149/mo. SheetEngine does web-grounded enrichment, email-pattern inference, and ICP scoring natively in the sheet you already work in — at flat BYOK pricing, with the total cost shown before you run.

100 founder seats · summer 2026

Stop renting credits.
Own the engine.

Founder pricing locks at $99/year — before the public launch price.