Relevance AI pricing: the Relevance AI cost per month, what an Action costs, and the full price ladder Relevance AI publishes in its docs but not on its pricing page.

How much does Relevance AI cost per month?

Relevance AI costs $0 a month on Free, $19 a month on Pro billed annually or $29 billed monthly, and $234 a month on Team billed annually or $349 billed monthly. Enterprise is custom and quoted by sales. Those four tiers are published in full at relevanceai.com/docs/get-started/pricing, read on September 8, 2026. Annual billing is described as saving 33% against the monthly rate, and on Team the arithmetic is exact: $234 against $349 is a 33% saving. On Pro, $19 against $29 works out slightly better than advertised at 34.5%.

What each tier includes matters more than the headline, because two counters run against every plan. Free gives you 200 Actions a month and a one time grant of 1,000 Vendor Credits, described as $2 of value, with 1 Build User, 1 Workforce, 1 Project and 30 days of task history. Pro adds 2,500 Actions a month and 10,000 Vendor Credits a month, 2 Build Users, unlimited Workforces, scheduled tasks, escalations to email and Slack, premium triggers for WhatsApp, LinkedIn and Telegram, and the ability to bring your own model keys. Team adds 7,000 Actions and 35,000 Vendor Credits a month, 5 Build Users and 45 End Users, 5 shared projects, calling and meeting agents, an analytics dashboard and priority support.

So the honest short answer for a solo operator is $19 a month, and for a team that needs more than two people touching agents it is $234 a month. The gap between those two numbers is twelve fold, and nothing in between exists. That is the single most useful thing to know before you start an evaluation on Pro with a team rollout in mind.

Why relevanceai.com/pricing shows no prices, and where the prices actually are

If you have arrived here after visiting Relevance AI's own pricing page and finding nothing you could budget from, you read it correctly. As of September 8, 2026 that page carries a single plan card labelled Enterprise, with Custom Actions and Custom Vendor Credits, a feature list and five links inviting you to talk to sales. It contains no dollar figure of any kind. There is no plan toggle to reveal, no annual switch hiding the ladder, and no geographic variant that shows more: we rendered the page with a headless browser from a United States address specifically to rule that out, because pricing pages routinely serve different content to a scraper than to a browser.

The full ladder is on the same domain, two clicks into the documentation, at relevanceai.com/docs/get-started/pricing. It is not a fragment or a leftover. It is a maintained, well built page with the four plans, per plan inclusions, a six section feature comparison covering platform, agent modes, app integrations, triggers, AI management, security and support, the top-up rates, the rollover rules and an FAQ. A second page at relevanceai.com/docs/admin/subscriptions/plans explains Actions and Vendor Credits and repeats the top-up pricing. Whoever maintains the documentation is doing a better job of pricing transparency than whoever owns the marketing site.

The Wayback Machine dates the change closely enough to be useful. Snapshots of the pricing page from March 2025, September 2025 and December 2025 all contain the ladder, and the December capture shows the current shape: $0, $19 and $29, $234 and $349, plus the $4,188 annual figure and the $840 Vendor Credit line. The March 2026 capture has lost almost all of it, the May 2026 capture retains only two stray figures, and the June, August and September 2026 captures contain no dollar amounts at all. Something changed between December 2025 and March 2026 and it has held for at least six months.

The practical consequence shows up in the search results. Run the query that brought you here and count how many of the page one results belong to Relevance AI: none of them do. The ranking set is aggregator guides and two competitors, 11x and Lindy, both of whom sell against Relevance AI. When a vendor stops answering a question about itself, the answer gets written by people with something to sell, and we include ourselves in that. Which is why every figure on this page names the page it came from and the day we read it.

Actions and Vendor Credits: what the two meters actually count

An Action is one run of a tool. Relevance AI's documentation defines it plainly: each time a tool runs it counts as an Action, whether that is sending a single email or executing a complex multi step workflow. Actions are charged when you run a tool directly, and when an agent or a workforce runs one on your behalf. And, in the same paragraph, a sentence worth reading twice: if the tool fails, this will still count as one Action.

A Vendor Credit is the cost of running the AI model, plus the cost of the tools you use. Relevance AI describes these as passed through at wholesale with no markup, and you can bypass them entirely by supplying your own OpenAI, Anthropic or Google keys, which is available on paid plans and not on Free.

Put the published numbers side by side and the difference between the two meters is stark. A Vendor Credit top-up is $20 per 10,000, which is $0.002 each. Pro includes 10,000 a month, which Relevance AI itself describes as $20 of value, and Team includes 35,000, described as $70. Divide either and you get $0.002. The included rate and the purchase rate are identical, which is exactly what no markup means and is unusually honest for the category.

Now do the same for Actions. A top-up is $80 per 1,000, or $0.08 each. Pro includes 2,500 in a plan that costs $19 a month. Attribute every cent of that $19 to Actions and pretend the $20 of Vendor Credits it also includes is worth nothing, which deliberately overstates the included price, and an included Action still comes out at $0.0076. The top-up rate is more than ten times that. On Team the same conservative calculation gives $0.0334 an Action against the same $0.08 top-up, a ratio of 2.4. Both ratios are real and the cheaper plan has the steeper cliff, which is the opposite of what most people assume.

This is not an accusation of anything. A vendor bundling a generous allowance into a low entry price and charging list for anything beyond it is ordinary commercial practice, and Relevance AI publishes both numbers where anyone can find them. The reason to write it down is forecasting. If you evaluate on Pro, run a light pilot month and multiply, your model will be wrong by up to a factor of ten in the direction that hurts.

The $349 coincidence, and why the third top-up is the mistake

Here is the arithmetic that changes a purchasing decision, using nothing but Relevance AI's own published rates. Actions must be topped up in increments of 1,000 at $80 a thousand, and top-ups are only available on a paid plan.

Take monthly billing first, because the numbers land exactly. Pro is $29 a month with 2,500 Actions. Add one top-up and you have 3,500 Actions for $109. Two, and it is 4,500 for $189. Three, 5,500 for $269. Four, and you have 6,500 Actions for $349.00. Team, billed monthly, is $349.00 and includes 7,000 Actions. The same money to the cent, and Team gives you 500 more Actions, 25,000 more Vendor Credits a month, three more Build Users, 45 End Users where Pro has none, calling and meeting agents, an analytics dashboard and priority support.

On annual billing the crossover arrives one step earlier. Pro is $19 with 2,500 Actions. One top-up is $99 for 3,500. Two is $179 for 4,500. Three is $259 for 5,500, and Team is $234 for 7,000. So the third top-up is where you begin paying more for less. The precise break-even is 2,688 extra Actions, but since you can only buy them in thousands, the practical rule is simple: two top-ups on annual Pro is still good value, and the third one is a signal to upgrade.

This matters because it runs against what the guides say. Several of the pages currently ranking for Relevance AI pricing describe topping up rather than upgrading as the flexible option that keeps costs down. On Relevance AI's own published rates, past the second thousand it is the option that costs more and delivers less. Nobody is being dishonest; it is the kind of thing you only notice if you sit down and multiply, and most roundups do not.

One nuance in Relevance AI's favour, and it is a real one. Purchased Action top-ups roll over to the next billing cycle even though plan Actions reset at renewal, and purchased Vendor Credits roll over indefinitely while you stay subscribed. So a top-up bought for one unusual month is not wasted if the month turns out lighter than expected. That softens the cliff for spiky workloads. It does not change the arithmetic for a workload that is simply larger than Pro.

Seats, Build Users and the part that decides whether Pro works for a team

Relevance AI splits people into two kinds and the distinction is easy to miss because both appear as single rows in a long comparison table. Build Users create and edit agents, tools and workforces. End Users are the colleagues you hand a finished agent to.

Free gives you 1 Build User. Pro gives you 2 Build Users and a dash under End Users, meaning none. Team gives you 5 Build Users and 45 End Users. Enterprise makes both unlimited. Alongside that, Free has 1 Project and 1 Workforce, Pro has 1 Project and unlimited Workforces, and Team has 5 shared Projects.

The consequence is worth stating plainly because it is the most common way a Relevance AI budget goes wrong. If your intention is to build agents and give them to a sales team to use, Pro cannot do that at any level of Action top-up, because top-ups buy capacity and not seats. The entry price for that shape of deployment is Team at $234 a month billed annually, not Pro at $19. If your intention is one or two operators building agents for their own work, Pro is well priced and the 2,500 Actions will go further than you expect.

There is one more capability boundary in the same place. Calling and meeting agents are a Team feature. If the reason you are evaluating Relevance AI is that you want an agent to sit on calls, Pro is not a smaller version of that product, it is a different one.

What we would use it for, and where a flat rate is the better buy

Relevance AI is a builder and a good one. If you want to assemble a multi agent go to market workforce, wire it into 2,000 or so applications, run it on your own model keys and keep the compute cost at wholesale, it does that well and the Vendor Credit policy is the most buyer friendly of any platform we have priced. Credits that never expire while you are subscribed, at cost, with bring your own key on every paid plan, is not marketing. It is a genuinely good deal and no other self-serve vendor in this comparison matches it.

The place it costs more than it looks is the Action meter on open ended, high volume work, and specifically on work that touches sources which fail. A prospecting flow that enriches from three providers is three Actions per record before the agent has written anything, and any of them that time out are billed the same as the ones that worked. At 5,000 records a month that is 15,000 Actions, which is two and a bit Team plans, or Team plus eight top-ups at $640. Volume is where a per run meter and a flat price diverge fastest.

The split we would make is between work that has a rule and work that needs judgement. Recurring, rule shaped jobs run the same way every week, and paying per run for them means paying again every week for an outcome whose shape you already know. That is the half WorkAgent is built for, at $149 a month flat, with no Actions to count, no second meter and no seat tiers, so a heavy month and a light month cost the same and a failed run costs nothing. Judgement work, one off investigations and genuinely novel builds are where a platform like Relevance AI earns its money, and we would not pretend otherwise.

If you are running that comparison properly, price both sides on your own numbers rather than on headline plans. Put your real monthly run count into the calculator above, including the runs you expect to fail, and compare the ceiling rather than the median, because the ceiling is what a bad month costs. We have done the same exercise on the neighbouring platforms: Gumloop pricing is the closest peer on shape, Lindy pricing is the closest on price, Copilot Studio pricing and Agentforce pricing cover the two enterprise meters, and our Relevance AI alternative page puts the two products head to head on what each is actually for.