Is Relevance AI Free? No Free Plan, and What Pro Costs

No. Relevance AI has no free plan today: its Free plan is retired for new signups and for existing accounts alike, so every organization needs Pro, Team or Enterprise before an agent can run ^[1]. The cheapest way in is Pro at $19 a month billed annually or $29 a month billed monthly, and that price covers the whole organization. Relevance AI's pricing pages describe no free trial. Bringing your own model keys trims one cost, not the subscription. Free use survives only at the edges.

The Free plan is gone, even for old accounts

The retirement reaches backward as well as forward. An organization that was created on Free has not been taken away: it still belongs to its owner, but nobody can work in it again until an admin upgrades it to Pro or Team. That matters if you signed up a while ago, built a few agents and assumed they were still idling on a free tier. They are not. Older blog posts and page templates still show a sign-up-for-free button, but they predate the change and no longer describe a real offer.

Finding the prices can also confuse a first-time visitor. Relevance AI's public pricing page shows only a custom Enterprise plan with a Talk to sales button; the self-serve Pro and Team prices are published in its pricing documentation, and you upgrade to them inside the app. So someone who lands on the public page may conclude there is no self-serve option at all. There is. The documentation lists two self-serve plans with fixed prices, and both are current.

The cheapest way in: what Pro covers

Pro is the realistic starting point for one or two builders testing an idea. It includes one project, unlimited agents, tools and Workforces, task scheduling, escalations and connections to 2,000+ apps. Paid annually, Pro comes to $228 a year by our arithmetic; paid monthly, twelve months come to $348, so committing up front saves $120 a year. For a short pilot, though, the monthly bill wins. By our arithmetic, billing monthly costs less than a year billed annually only if you leave within seven months. From the eighth month on, annual billing is cheaper.

Pro also has hard edges. Only two users can be added to a Pro organization, and Pro includes no End Users at all, while Team includes 45. Calling and meeting agents, the analytics dashboard and bulk scheduling start on Team, and SSO, RBAC, audit logs and agent evaluations are Enterprise-only. Team costs $234 a month billed annually or $349 a month billed monthly. That jump buys people and features more than volume: measured per included Action, Pro is by far the cheaper plan.

Why your own model keys don't make it free

Relevance AI bills on two meters, and a free path would have to dodge both. Actions count Tool runs: one run is one Action whether it sends a single email or runs a many-step workflow, and a failed run still counts, as do retries. Vendor Credits pay for the AI model and tool costs, passed through at wholesale with no markup, and 10,000 of them cost $20 as a top-up pack. Plan Actions reset at each renewal, while Vendor Credits roll over indefinitely as long as you stay subscribed on Pro or Team.

Bringing your own keys for OpenAI, Anthropic, Google and other providers is optional on every plan ^[2]. It moves model cost off Vendor Credits and onto your provider's bill, keeping the credits as a fallback. It does nothing to the other meter. Every Tool run still spends an Action, the subscription stays the same, and the Terms say Relevance AI's fees are not reduced for usage on your own keys. A team that pays its model provider directly still pays for its plan and for every Action beyond the allowance.

That allowance is where a cheap start can quietly grow. Pro includes 2,500 Actions a month (30,000 a year), and an organization admin can buy more as a top-up pack of $80 per 1,000 Actions; by our arithmetic that is $0.08 an Action. Pro billed monthly plus one 1,000-Action top-up comes to $109 for a month with 3,500 Actions, by our arithmetic.

If your model spend fits Pro's 10,000 Vendor Credits a month, Team billed monthly ($349 a month) is cheaper than Pro billed monthly plus Action top-ups above 6,500 Actions a month on monthly billing, or about 62,000 a year (roughly 5,200 a month) on annual billing, by our arithmetic; heavier model use lowers that crossover. Action top-ups roll over to the next billing cycle; Vendor Credit top-ups, $20 per pack of 10,000, last indefinitely while you stay subscribed.

What you can still use without paying

A few pieces cost nothing, though most lead back to a plan. The MCP server at mcp.relevanceai.com and the Claude Code plugin are free, so coding agents such as Claude Code, Codex or Cursor can build and run Relevance AI agents from your editor ^[3]. The catch sits on the same docs page: agent runs and tool executions are billed per your plan. The agent-skills repository on GitHub, which teaches coding assistants the 46 MCP tools, adds context for those assistants without adding a plan of its own.

The Marketplace is the other near-free corner. Free listings are marked Free and can be cloned again whenever you like, and the docs point to pre-built agents there such as the HubSpot Assistant, the Gmail Assistant and the Calendar Bookings Scheduler. Cloning is not running, however. The Marketplace docs say a new customer needs a Pro or Team subscription to run cloned agents, and a paid listing is an extra purchase on top ^[4]. The documentation is public, and the community forum is open on every plan.

Learning material is the last piece. Relevance AI Academy offers structured courses that walk you through building on the platform step by step, which helps you plan an agent before the first billed run. If you decide to go ahead, our Relevance AI tutorial covers the first build on Pro, from describing a process to Invent through cloning a template and adding a trigger, with approval mode on while you test. Reading first is cheap. Running is not.

No open-source or self-hosted route

Relevance AI runs as a hosted service; your organization's data is stored in the region (US, EU or Australia) chosen when the organization is created, and the docs describe no self-hosted edition. The builder itself runs in a desktop browser, and the mobile apps cover Chat only. There is no package to download, no server to run and no community edition to fall back on when the budget is zero. For readers searching for an open-source version of Relevance AI, the honest answer is that none is described.

If free or self-hosted is a hard requirement, look at different tools. n8n offers a free self-hosted Community Edition with no execution quota, though it leaves out projects, SSO, workflow and credential sharing and Git version control, which need a paid license; its Cloud plans open with a 14-day trial that needs no credit card. CrewAI publishes an MIT-licensed open-source framework, and its AMP platform has a free Basic tier capped at 50 workflow executions a month. Both lean on workflows or code more than a hosted agent canvas. Our Relevance AI alternatives guide compares them by need.

Keeping the first paid month predictable

Payment is by credit card only (no debit cards), and Relevance AI may add a surcharge on card payments. Usage alerts, which send an email at a level you set, and usage limits, which stop usage at a hard ceiling, work on every plan. Set both before your first agent goes live. Spend Controls, the automatic top-ups on Pro and Team, have no cap on how many top-ups they buy, so keep them off until you know your volume. The docs name running out of Actions or Vendor Credits as the most common cause of agent failure, so a hard limit trades that risk for a predictable bill.

Two more habits protect a small budget. An embedded or cloneable agent is public, and anyone with the link runs it on your credits, so leave sharing off during a pilot. Canceling is done in the app by an organization admin and takes effect at the end of the current paid cycle; fees are non-refundable for change of mind, and unused Actions and Vendor Credits expire when you cancel. Our Relevance AI pricing breakdown covers Team, top-ups and the break-even points in full.

What the cheapest real Relevance AI start costs

There is no free floor, so the real question is how small a paid start can be. The answer is Pro at $19 a month billed annually or $29 a month billed monthly, shared by one organization with two Build Users and one project. Model costs come out of the included Vendor Credits or move to your own provider's bill, while every Tool run draws on the Action allowance either way. Pay monthly while you test, switch to annual once an agent earns its keep, and set a usage limit on day one.