Licensing

Licensing models for Autopilot under Unified Pricing and Flex, Basic, Plus, and Pro model access and project usage expectations, and model and project classification methodology.

Autopilot is available under two licensing models: Unified Pricing and Flex. Both models draw from the AI and Agentic usage pool, the monthly limit shared across UiPath AI and Agentic products. Autopilot is not available on the Free plan.

Autopilot shares its pool-sizing model and its no-fallback behavior with Delegate and Cartographer: the monthly limit scales the same way by license tier under both Unified Pricing and Flex, and all three pause instead of drawing from paid units once the limit and any top-up are spent. For that shared model, including the license-tier sizing tables for both plans, refer to Usage pool consumption for Autopilot, Delegate, and Cartographer.

Autopilot actions (such as Fix Expression, Generate Expression, and Generate Sequence Name) are not subject to usage pool consumption and remain available without limits. For the full list of actions, refer to Autopilot actions.

Unified Pricing

Autopilot consumption depends on your organization's plan (Enterprise, Community, Basic, or Trial) and the user license assigned (Basic, Plus, Pro, App Tester, or App Test Developer). For the pool size each combination receives, refer to Usage pool consumption for Autopilot, Delegate, and Cartographer.

Flex

Autopilot consumption depends on your organization's plan (Enterprise, Community, or Trial) and the user license assigned (Attended User, Citizen Developer, Tester User, or Automation Developer). For the pool size each combination receives, refer to Usage pool consumption for Autopilot, Delegate, and Cartographer.

What to expect with Basic, Plus, and Pro licenses

Model access is separate from monthly capacity:

• Basic users can access all available cost-efficient and standard models in their tenant through AI Trust Layer, such as Kimi K2.7, Terra, and Claude Sonnet. Basic users cannot access premium models, such as Claude Opus, Claude Fable, and GPT Sol.
• Plus and Pro users can access all models made available in their tenant through AI Trust Layer, including premium models.

Note:

Model availability may differ from the standard offering depending on the tenant configuration and deployment region. For details, refer to AI features and model routing.

The following table summarizes the estimated number of projects each monthly usage pool can support. The values are alternatives, not cumulative — each estimate shows what the full pool can support if spent on that activity alone. For example, a Basic pool can build 4–5 high-complexity projects, or build 10 medium-complexity projects, or fix design-time errors in up to 10 high-complexity projects, or refactor up to 10 high-complexity projects.

License tier	Build from scratch	Fix design-time errors	Refactor
Basic — cost-efficient models (e.g. Kimi K2.7)	10 medium-complexity projects or 4–5 high-complexity projects	Up to 10 high-complexity projects	Up to 10 high-complexity projects
Plus — standard models (e.g. Terra, Sonnet)	12–15 medium-complexity projects	20–25 medium-complexity projects	35–40 medium-complexity projects
Plus — premium models (e.g. Opus, Sol)	5–6 high-complexity projects	7–9 high-complexity projects	8–10 high-complexity projects
Pro — standard models (e.g. Terra, Sonnet)	45–50 medium-complexity projects or 30–40 high-complexity projects	90–100 medium-complexity projects or 40–50 high-complexity projects	150–180 medium-complexity projects or 125–135 high-complexity projects
Pro — premium models (e.g. Opus, Sol)	18–20 high-complexity projects	20–25 high-complexity projects	30–35 high-complexity projects

Note:

• These estimates apply to individual projects. For solutions composed of multiple interdependent Complex projects, aggregate consumption depends on the number of projects and the extent of cross-project work (such as building or refactoring the solution as a whole) and may significantly exceed any single per-project estimate. Plan monthly usage capacity at the aggregate solution level rather than per project.
• Estimates are not fixed limits. Actual capacity varies with the tasks Autopilot performs, the model used, and the complexity of each project. Because model choice affects usage, the project counts do not scale directly with the usage-pool multipliers.
• The table uses Unified Pricing license names. For the equivalent Flex licenses, refer to Usage pool consumption for Autopilot, Delegate, and Cartographer.
• For model categories, relative cost ratings, and the definitions of Simple, Medium, and Complex projects, refer to Methodology for usage estimates.

Autopilot for Everyone

Autopilot for Everyone has its own licensing model: activities are measured through Autopilot actions, with one action counted from the time you select a button or write a prompt until you receive a result.

Examples of Autopilot actions include:

• Generate tests in Test Manager.
• Generate code in Studio Desktop when Autopilot creates coded workflows or generates XAML workflow files.
• A successful paste in Clipboard AI.

For details, refer to Autopilot for Everyone licensing.