Clay Pricing 2026: Which Plan You Actually Need
•
Most seed to Series B teams should start Clay on Launch at $185/month and move to Growth at $495/month the month they need CRM sync. Billed annually that is $2,004 or $5,352 a year. The Free plan is a demo, and Enterprise is not a conversation you need to have yet.
The harder question is what you will actually spend, because Clay bills on usage. Two teams on the same plan can pay ten times more per enriched record than each other. Here is how to pick a plan, what a year really costs, and the three mistakes that double the invoice.
Which Clay plan you need
Where you are | Plan | Cost per year |
|---|---|---|
Testing Clay, under 40 records/month | Free | $0 |
Founder-led outbound, no CRM sync, under 400 records/month | Launch | $2,004 |
First GTM or RevOps hire, syncing to your CRM | Growth | $5,352 |
Several motions running, 1,000+ records/month | Growth, expanded credits | $5,352+ |
The line that matters is CRM sync. It is the single feature that forces the jump from $185 to $495, and there is no way around it: Launch cannot write to Salesforce, HubSpot, or Attio natively. If your reps live in the CRM, you are on Growth, and paying for Launch first is a false economy.
If you are pre-CRM and running lists into a sequencer, Launch covers it comfortably. Most founders I talk to over-buy here.
Clay pricing calculator
Now put your own numbers in. Set your plan, how many records you enrich each month, how deep you enrich them, and which AI model you run. It returns your annual cost, your real cost per record, and how much of your allowance you are actually burning.
It tracks both meters, because Clay bills two. Data Credits run out first for most enrichment work. Actions run out first if you sync or export the same records on a schedule, and Actions cannot be topped up, so that ceiling means a plan upgrade rather than a credit purchase.
The fastest thing to try: change the AI model dropdown. Most teams set a model once during setup and never look at it again, and the spread between Clay's cheapest and most expensive model is 37 times per run.
If the numbers came out worse than you expected, the rest of this guide explains what drives them and what to do about it.
Clay pricing plans in 2026
Clay restructured its pricing on 11 March 2026, replacing the Starter, Explorer, and Pro tiers with two self-serve plans. Current numbers from Clay's pricing page:
Plan | Monthly | Annual (per mo) | Data Credits/mo | Actions/mo |
|---|---|---|---|---|
Free | $0 | $0 | 100 | 500 |
Launch | $185 | $167 | 2,500 | 15,000 |
Growth | $495 | $446 | 6,000 | 40,000 |
Enterprise | Custom | Custom | Custom | Custom |
Every plan includes unlimited seats, so team size does not change your bill. Annual billing saves just under 10% and delivers credits upfront. Clay does not publish an Enterprise price.
One thing worth knowing before you compare tiers: at base allowance, Growth costs slightly more per Data Credit than Launch, not less. The extra $310 a month buys CRM sync, HTTP API, webhooks, and Ads. It does not buy cheaper data. Both plans let you expand the credit allowance without changing tier, and the per-credit rate falls as you do.
The number that actually decides your bill
Clay charges Data Credits for the data you buy and Actions for the work it performs. Actions rarely bind for a team at your stage. Credits do, and how many a record costs is almost entirely a build decision.
Here is the same contact record built two ways, using unit costs from Clay's own documentation.
Step | Lean build | Heavy build |
|---|---|---|
Import and source a list | 0 | 0 |
LinkedIn profile | 0.5 | 0.5 |
Work email | 0.5 (one provider) | 1.0 (waterfall, two hit) |
Email validation | 0 (free provider) | 0 |
Mobile phone number | not used | 5 (People Data Labs) |
AI research | 1 (Claygent Helium) | 3 (Claygent Argon) |
AI copy | 0.4 (GPT-5 Mini) | 1.5 (Claude 4.5 Sonnet) |
Credits per record | 2.4 | ~11 |
Cost per record on Growth | $0.20 | $0.91 |
Sourcing lists, importing your own data, formulas, and CSV exports are all free. Failed lookups are free too, so a waterfall that finds nothing costs nothing.
Phone numbers are priced by provider, not by data point, so pick carefully: the same mobile number runs about 5 credits on People Data Labs against 25 on Datagma.
Clay's guidance says a fully-enriched record runs 6 to 20 credits. That range quietly assumes you are buying phone numbers and running premium AI. Plenty of GTM motions at seed and Series A need neither.
Practical version: on Growth, a lean build gets you about 2,500 records a month inside your allowance. A heavy build gets you about 545.
The cheapest lever nobody pulls
Clay's AI pricing docs publish a per-model credit cost that I have not seen reproduced anywhere else. The spread is bigger than most people expect.
Model | Credits per row (content) |
|---|---|
GPT-5 Nano | 0.2 |
GPT-5 Mini | 0.4 |
GPT-4o | 1 |
Claude 4.5 Sonnet | 1.5 |
Claude 4.6 Opus | 7.5 |
Opus costs 37 times more per row than Nano. Across 2,000 rows a month that is $1,238 versus $33, decided by one dropdown that most teams never revisit after setup.
You do not need a frontier reasoning model to draft a first line of outbound copy. Test the cheap model on 10 rows, and only trade up if the output actually fails. This is the fastest saving available in Clay, and it takes about two minutes.
Why most teams overpay
Here is the part that changes budgets, and it has nothing to do with which providers you chose.
Your subscription is sunk whether you burn the credits or not. So your real cost per record is the plan price divided by the records you actually enriched.
Records enriched/month | Real cost per record on Growth |
|---|---|
100 | $4.95 |
200 | $2.48 |
400 | $1.24 |
1,000 | $0.49 |
A team enriching 100 records a month on Growth pays $4.95 per record. The data inside those records cost about $0.49. They are overpaying tenfold, and no amount of provider tuning fixes it, because the leak is not in the waterfall. It is the 90% of the allowance that expired unused.
This is the most common Clay cost problem I see at seed to Series B, and it is almost always a symptom of something else. Enrichment runs on lists someone builds by hand every other week instead of firing on a trigger. Or the table was built for a campaign that ended two quarters ago and nothing replaced it.
This is the utilisation figure the calculator returned for you. Check it against Settings, then Credit usage, before you upgrade, downgrade, or churn. Under 60% on a normal month means you have a workflow design problem wearing a pricing costume.
Three ways to avoid a surprise invoice
Do not buy top-ups every month. Running out mid-cycle lets you buy more credits at a 30% premium. That is fine for a one-off push. If it happens two months running, expand your credit tier inside your current plan instead, which carries no premium and is the option most people forget exists.
Watch Actions if you sync a lot. Every record written to your CRM, sequencer, or Slack costs one Action and zero credits. Growth includes 40,000 Actions, which is comfortable for enrichment work but tightens fast if you are syncing a large CRM on a schedule. Actions cannot be topped up, so hitting that ceiling means a plan upgrade.
Know what rolls over. Data Credits roll over up to twice your monthly allowance on monthly plans, and 15% carries at renewal on annual plans. Actions expire every cycle, full stop.
What Clay costs beyond the subscription
The subscription is the part you can plan for. Three Clay-specific costs sit underneath it, and only one is optional.
Re-enrichment never stops. Contact data decays 25% to 30% a year, mostly through job changes. A 6,000-record database therefore needs roughly 1,700 records refreshed annually just to stay accurate. At 6 credits each that is close to 10,000 credits, or a month and a half of a Growth allowance spent re-buying data you already own. Almost nobody budgets for it, and it is what turns a comfortable plan into a tight one around month four.
The leaks are documented and avoidable. Clay names four common causes of unexpected credit deduction: auto-update refreshes firing enrichments on a schedule, AI columns charging on every row they process, re-running a column charging again from scratch, and cancelled runs still billing for requests already sent. Every one is a settings problem rather than a pricing problem.
Someone has to own it. The cost that never appears in a comparison table is the hours spent designing tables, ordering waterfalls, and watching the credit dashboard. That is the difference between $0.20 and $4.95 per record.
Cost | On the pricing page | Recurring |
|---|---|---|
Subscription | Yes | Yes |
Top-ups at a 30% premium | No | Only if under-sized |
Re-enriching stale records | No | Yes, permanently |
Auto-update and duplicate burn | No | Until you fix it |
Actions ceiling forcing an upgrade | No | Step change |
The person who builds it | No | Yes |
One sequencing note, because it decides all of the above: buy the CRM before you buy Clay. Enriching records with nowhere to put them is how teams end up paying for data months before anything acts on it.
The useful way to read this table: your enrichment spend only pays back if the enriched record reaches the CRM and changes what a rep does. Everything above the CRM line is cost until that happens.
What changed in March 2026
Half the Clay pricing guides online still quote retired numbers, so if you are comparing sources, check for these four changes.
Starter ($149), Explorer ($349), and Pro ($800) were replaced by Launch and Growth, and the window to switch between legacy tiers closed on 10 April 2026. Marketplace data got 50% to 90% cheaper, which Clay explains candidly in its published pricing memo. Failed lookups stopped consuming credits. And CRM sync moved down from the $800 tier to the $495 one, which is the change that most improves the maths for a startup.
Verdict: is Clay worth it at seed to Series B?
Yes, if someone owns it. Clay is meaningfully better value than it was in February, and at $2,004 to $5,352 a year it is cheap next to the SDR headcount it displaces.
The catch is that Clay is a build, not a subscription. Two teams on the identical Growth plan can pay $0.20 or $4.95 per enriched record, and the gap is entirely in how the tables were made: whether waterfalls stop at the first valid result, whether you reached for Opus when Nano would do, whether enrichment fires on a trigger, and whether the enriched record lands somewhere your reps actually use it.
That last one decides whether any of this was worth it. A record that costs $0.20 to enrich and then sits in a Clay table nobody opens cost you $0.20 for nothing. Getting the Clay to CRM handoff right is what turns enrichment spend into pipeline, and it is the step most teams bolt on last.
If you want to go deeper on the build side, start with the complete Clay guide.
Frequently asked questions
How much does Clay cost?
Clay costs $185 per month on Launch and $495 per month on Growth, or $167 and $446 on annual billing. That is $2,004 to $5,352 a year. The Free plan includes 100 Data Credits a month. Enterprise pricing is custom and Clay does not publish it.
Which Clay plan should a startup choose?
Launch at $185 covers founder-led outbound up to roughly 400 records a month. Growth at $495 is the first tier with native CRM sync, so any team running its process through Salesforce, HubSpot, or Attio needs it. Team size does not affect either price.
How many Clay credits does one enriched contact use?
A lean build using free validation and a cheap AI model costs about 2.4 Data Credits, roughly $0.20 on Growth. Adding a mobile number and a frontier reasoning model pushes it past 11. Clay's own guidance of 6 to 20 credits assumes the heavier end.
How does Clay pricing work?
Clay charges by usage, not by seat, so all plans include unlimited users. You draw from two pools: Data Credits, which buy third-party data from Clay's marketplace, and Actions, which meter platform work like enrichments, AI runs, and exports to your CRM.
Do Clay credits roll over?
Data Credits roll over on monthly plans up to twice your monthly allowance. Annual plans carry over 15% of unused credits at renewal if you renew at the same tier or higher. Actions never roll over and cannot be topped up separately.
Is Clay worth it for a small team?
At $2,004 to $5,352 a year it is inexpensive next to the headcount it replaces, but only if someone owns the build. Clay rewards teams who design workflows deliberately and punishes ones who set it up once and stop looking at the credit dashboard.
Sparsh Gupta, Founder of Automation Jinn and an Official Attio Expert Partner, helps seed to Series B B2B SaaS teams build Clay systems that spend less per record and put every enriched field somewhere their reps actually use it. If you want the right plan sized and the workflows built properly the first time, book a discovery call.
