CrewAI Pricing and Plans: Basic Free Limits, Enterprise Custom Tiers, and Third-Party Checks (2026)
Compare CrewAI Basic free limits, Enterprise tiers, and MIT self-hosting against the RelayX Professional listing and pre-purchase checks.

Open article contents
Reviewed on November 14, 2026. Estimating CrewAI costs starts by separating the MIT open-source framework, CrewAI's hosted platform tiers, and third-party account listings. On its current public pricing page, CrewAI lists a free Basic tier and custom Enterprise plans, whereas the RelayX CrewAI listing still shows a 30-day "CrewAI PROFESSIONAL" entry. Because public tier names and availability differ, verify the exact conditions before buying.
Official plans and hosted execution limits
| Official option | Public price | Listed execution quota and key capabilities |
|---|---|---|
| Open-source framework (MIT) | US$0 (Free) | Free to run locally or self-host; you pay underlying LLM API token costs separately |
| Basic (Hosted platform) | US$0 (Free) | 50 workflow executions per month, up to 2 automations, Studio visual editor, GitHub integration, and export as an MCP server or UI component |
| Enterprise | Custom | Execution volume sized to workflow needs; adds SSO (Microsoft Entra, Okta), RBAC, PII redaction, cloud or private VPC deployment, and 45-day onboarding |
Source: the official CrewAI pricing page, reviewed on November 14, 2026. Notice that CrewAI's public page currently centers on Basic and Enterprise rather than a standalone self-serve paid mid-tier. In addition, even on the free Basic hosted plan, third-party LLM provider token costs (such as OpenAI or Anthropic) are billed separately from CrewAI.
Self-hosting the framework vs. using the hosted runtime
If your team has basic Python engineering and server operations support, deploying the MIT open-source CrewAI library on your own infrastructure avoids platform subscription fees and removes the 50-execution monthly cap on Basic. Your ongoing expense is essentially whatever LLM API tokens your agent runs consume.
By contrast, CrewAI's hosted runtime makes sense when teams want Studio visual editing, managed tracing, cron scheduling, or quick export as an MCP server without maintaining deployment infrastructure. Start on Basic with one or two representative workflows to measure actual token usage and run frequency before negotiating Enterprise terms or moving to self-hosted orchestration.
How the RelayX listing differs
At review time, RelayX lists a 30-day "CrewAI PROFESSIONAL" account item at CNY219 (reference price CNY269) with a Coming soon sales status, meaning direct checkout is not yet open.
Because CrewAI's current public pricing page no longer displays an open self-serve Professional tier, do not assume a third-party "PROFESSIONAL" label maps to a published monthly quota. Once sales open, check four details explicitly:
- Which official account tier or legacy workspace entitlement the delivered account actually holds.
- What monthly workflow execution limit and automation count apply in practice.
- Whether the package covers only CrewAI platform access or includes (or still requires your own) LLM API keys.
- How workspace isolation, API-key security, and replacement guarantees are handled if an account is shared or restricted.
Pre-purchase checklist
- Check whether self-hosting the free MIT framework with your own LLM API key already satisfies your workload.
- If testing the hosted platform, confirm whether 50 monthly executions and 2 automations on Basic cover your pilot.
- Verify that the third-party listing has moved out of Coming soon and review its live delivery description before paying.
- Budget platform fees and LLM API token bills separately to avoid underestimating total operating cost.
After choosing your deployment model, review the current RelayX CrewAI conditions separately from official public tiers rather than relying solely on a historical tier name.


