Mas você também pode fornecer outros `servers` alternativos, por exemplo, se quiser que a *mesma* interface de documentação interaja com ambientes de staging e produção.

Se você passar uma lista personalizada de `servers` e houver um `root_path` (porque sua API está atrás de um proxy), o **FastAPI** inserirá um "server" com esse `root_path` no início da lista.

Por exemplo:

it.  And you must show them these terms so they know their rights.

  We protect your rights with a two-step method: (1) we copyright the
library, and (2) we offer you this license, which gives you legal
permission to copy, distribute and/or modify the library.

  To protect each distributor, we want to make it very clear that
there is no warranty for the free library.  Also, if the library is

          exercise the Licensed Rights in the Licensed Material to:

            a. reproduce and Share the Licensed Material, in whole or
               in part; and

            b. produce, reproduce, and Share Adapted Material.

       2. Exceptions and Limitations. For the avoidance of doubt, where
          Exceptions and Limitations apply to Your use, this Public

    /*
     * Equivalence's type argument is always non-nullable: Equivalence<Number>, never
     * Equivalence<@Nullable Number>. That can still produce wrappers of various types --
     * Wrapper<Number>, Wrapper<Integer>, Wrapper<@Nullable Integer>, etc. If we used just
     * Equivalence<? super T> below, no type could satisfy both that bound and T's own

 * {@code 0x7F}), and to strings containing such characters.
 *
 * <p>ASCII utilities also exist in other classes of this package:
 *
 * <ul>
 *   <!-- TODO(kevinb): how can we make this not produce a warning when building gwt javadoc? -->
 *   <li>{@link StandardCharsets#US_ASCII} specifies the {@code Charset} of ASCII characters.
 *   <li>{@link CharMatcher#ascii} matches ASCII characters and provides text processing methods

//
// `GetID` returns a unique id for the wrapped tensor which is used to maintain
// a map (`tensorflow::eager::TensorTape`) from the wrapped tensor to the id of
// the op that produced it (or -1 if this tensor was watched using
// `GradientTape::Watch`.) The op_id is simply a unique index assigned to each
// op executed under the tape. A separate map (`tensorflow::eager::OpTape`)

          <description>The identifier for this artifact that is unique within the group given by the
            group ID. An artifact is something that is either produced or used by a project.
            Examples of artifacts produced by Maven for a project include: JARs, source and binary
            distributions, and WARs.</description>
          <type>String</type>
        </field>
        <field>

  }

  // This entry point is used by the other create method but also directly by
  // ImmutableMap.Builder, so that it can remember any DuplicateKey encountered and produce an
  // exception for a later buildOrThrow(). If builder is null that means that a duplicate
  // key will lead to an immediate exception. If it is not null then a duplicate key will instead be

          per_device_output_tensors.rbegin()->size() != first_op_output_count) {
        first_bad_status.reset(TF_NewStatus());
        TF_SetStatus(first_bad_status.get(), TF_INTERNAL,
                     "Parallel ops produced different numbers of tensors.");
      }
    }
  }
  if (first_bad_status != nullptr) {
    TF_SetStatus(status, TF_GetCode(first_bad_status.get()),
                 TF_Message(first_bad_status.get()));

* 100% do código utiliza <abbr title="Type annotations do Python, com isso seu editor e ferramentas externas podem te dar um suporte melhor">type annotations</abbr>.
* Usado para aplicações em produção.

## Recursos do Starlette