Clay
A table where every row decides which provider to call, in what order, and stops as soon as one answers.
Clay is a spreadsheet that runs data operations. Each column is a call to a provider, an AI agent or your own API, and each row runs the logic on its own: try the cheap provider, and only if it fails call the expensive one. It is the reference tool for multi-provider waterfalls, with a free plan and paid plans at 185 and 495 dollars a month.
What it is for
- Design the waterfall before you build itTrying providers in order, per row, with a stop at the first hit is exactly what a chain should do. Doing it visually first tells you what your own chain must reproduce.
- Call anything a provider does not offerHTTP columns and agent columns mean the row can call your own API, read a page, or ask a model, without waiting for an integration to exist.
- Enrich what the CRM is missingOn the paid tiers it syncs both ways with the CRM, so the table becomes the place where an account gets completed rather than a file someone re-imports.
What I measured
- Their free plan is not a demo: unlimited seats and tables, multi-provider waterfalls, 100 data credits and 500 actions a month, capped at 200 rows per table. Enough to design a chain and prove it works before paying anything.
- Data credits and actions are two separate meters, and confusing them is how a plan runs out mid-run. Launch includes 2,500 credits and 15,000 actions, Growth 6,000 and 40,000.
- Their search and enrichment costs are billed whether or not a row is found, unlike the specialists that bill on result only. On a list with a low hit rate, that difference is the whole argument for putting a cheap specialist in front.
What it is not for
- You run one stable enrichment sequence at volume. At that point the flexibility costs more than it returns and a direct API chain is cheaper.
- You want a tool your SDRs open themselves. This is an operator surface, not a rep surface.
- You are looking for a sender. It builds and enriches lists, it is not where the campaign runs.
What it costs, as the vendor publishes it
| Free | 0 dollars | 100 data credits and 500 actions a month, 200 rows per table |
|---|---|---|
| Launch | from 185 dollars a month | 2,500 data credits and 15,000 actions a month |
| Growth | from 495 dollars a month | 6,000 data credits and 40,000 actions a month |
| Enterprise | custom, annual commitment | quoted with the contract |
- Two meters, not one: data credits pay for provider calls, actions pay for the steps around them.
- Phone enrichment and signal tracking start on the paid tier, not on the free one.
- Unlimited seats on every plan, including the free one.
Vendor list price, read on their pricing page on 18 Sep 2026. Their pricing page. Read 18 Sept 2026.
No video on this one yet. When there is one it goes here, recorded on the real account. The page says so either way.
Questions people ask
Is the free plan usable?
For design, yes. Unlimited seats and tables, waterfalls included, 200 rows per table. It is enough to prove a chain works before paying for volume.
What is the difference between a data credit and an action?
A data credit pays for a provider call that returns data. An action pays for a step in the table. They run out at different speeds and only one of them is usually watched.
Is it cheaper than calling providers directly?
No, and it does not claim to be. You pay for orchestration and for the ability to change your mind. At a stable high volume, a direct chain wins on cost.
Should an SDR use it?
No. It is an operator tool. What reaches a rep should be a finished list, not a table with a hundred columns.