Pricing
Free trial, then metered.
Try Canoa free — sign up takes one minute, and you get enough usage to spec out a real project before we ask for a card. Once you've used it enough to know, add a card and continue at $4 per million tokens. No subscription, no minimum, no seat license. Stop using it = stop being billed.
Always-on watching is on for everyone from day one — no separate tier, no upgrade gate. It just counts against your token usage like a chat would.
Free trial
$0
Enough usage to spec out a real project — call it ~1.5 million tokens, or a few dozen substantive conversations with the FF&E specialist plus a week or two of always-on watching on a typical sheet.
- Sign up via Google. One OAuth popup, no password, no payment.
- All skills in the toolkit (sourcing, scheduling, sustainability, planning, presentation)
- Master schedule Sheet template + import from past project / Pinterest / fresh
- Always-on Sheet-watching from day one — no separate tier
- Email-on-find via Sheets comment mentions
After the trial · per million tokens
$4/M tokens
Add a card via Stripe when the trial ends. Pay only for the tokens we actually process. Big sheet, big bill; small sheet, small bill. Cancel = stop using = $0.
- Everything in the trial, unlimited
- Always-on watching, daily or sub-daily cadence
- Multiple sheets per account
- Cache hits cost ~nothing — Canoa's catalog gets smarter with every user, lowering your effective rate over time
- Real-time usage dashboard
- Billed monthly via Stripe based on actual usage. No subscription.
What does usage look like in practice?
A token is roughly 0.75 words. Canoa's skills read your sheet rows, process them through Claude, and write findings back. The bigger or more frequent the work, the more tokens — and the more your bill. Below is the realistic shape for typical projects (assumes the full skill suite running plus daily always-on watching; numbers vary with column density, cache-hit rate, and which checks fire).
| Project size | Cadence | Tokens / month | Cost |
|---|---|---|---|
| 50 rows | weekly watch + occasional chats | ~200K | $0 (trial) |
| 200 rows | weekly watch + active chats | ~1M | $0 (trial) |
| 200 rows | daily watch + active chats | ~6M | ~$18/mo |
| 1,000 rows | weekly watch + occasional chats | ~4M | ~$10/mo |
| 1,000 rows | daily watch + active chats | ~30M | ~$114/mo |
| 5 sheets × 1,000 rows | daily | ~150M | ~$594/mo |
Solo designers with one project at a time will likely never see a bill before the project closes. Studios with multiple active projects and daily watching pay in proportion to the work that actually happened. Numbers above are rough — real usage depends on cache-hit rate (the catalog gets smarter, so the same workload costs less over time).
Why metered, not subscription
Different studios have wildly different sheet sizes and project cadences. A flat subscription would punish solo designers (who barely use any tokens) and underprice big studios (who run multi-sheet, multi-project, daily watching). Per-token maps your bill to actual cost. Anthropic charges Canoa per token; we mark up to 2× and pass through. The Canoa catalog cache lowers Canoa's effective cost on cache hits — which over time lowers the effective rate you pay too, because cache hits are billed at the same rate but cost us less to serve, and we plan to drop the rate as the catalog matures.
How Canoa keeps the bill low
Two mechanisms. Cache: every product Canoa enriches once is cached for everyone — when another designer asks about the same SKU, the response comes from cache, often without an LLM call at all. Model selection: Canoa picks the cheapest Anthropic model that can do each check correctly — Haiku for routine validation (URL checks, missing-data flags), Sonnet only when the check requires deeper judgment. That keeps your token cost meaningfully lower than always running the most expensive model.
What about studio + enterprise?
Studios with very high volumes (sustained 100M+ tokens/month), custom checks, or compliance requirements (SSO, on-prem dataset, audit logs) — get in touch. Volume discounts kick in above 100M tokens/month.