A role or user in AWS is rarely governed by a single policy — an identity-based policy, one or more attached managed policies, and inline policies can all apply at once, and the actual set of statements that IAM evaluates is the union of all of them. IAM Effective Permissions Calculator takes up to three pasted policy JSON documents and combines them the way IAM combines multiple attached policies: every Allow statement from every document is pooled into one Allow list, and every Deny statement is pooled into one Deny list, each tagged with which policy it came from.
This is a deliberately narrow, honest tool: it performs a syntactic union, not a full request simulation. It will not tell you whether s3:GetObject on a specific bucket is actually allowed once conditions, explicit Denies, and overlapping wildcards are all taken into account — that requires evaluating one specific request against the combined statement set, which is what the separate IAM Policy Simulator tool does. What this tool gives you instead is the raw combined picture: everything that grants access, and everything that explicitly denies it, laid out side by side across all the policies you attached.
That combined view is still useful on its own: it is the fastest way to see, across several policies at once, how many Allow statements exist in total, whether any explicit Deny statements are present at all (a common source of surprising access restrictions), and which policy each statement originated from — before diving into a full simulation of a specific request.