The Manifest Format

The [package] section

The first section in a Cargo.toml is [package].

name = "hello_world" # the name of the package
version = "0.1.0"    # the current version, obeying semver
authors = [""]

All three of these fields are mandatory.

The version field

Cargo bakes in the concept of Semantic Versioning, so make sure you follow some basic rules:

The build field (optional)

This field specifies a file in the repository which is a build script for building native code. More information can be found in the build script guide.

# ...
build = ""

The exclude and include fields (optional)

You can explicitly specify to Cargo that a set of globs should be ignored or included for the purposes of packaging and rebuilding a package. The globs specified in the exclude field identify a set of files that are not included when a package is published as well as ignored for the purposes of detecting when to rebuild a package, and the globs in include specify files that are explicitly included.

If a VCS is being used for a package, the exclude field will be seeded with the VCS’ ignore settings (.gitignore for git for example).

# ...
exclude = ["build/**/*.o", "doc/**/*.html"]
# ...
include = ["src/**/*", "Cargo.toml"]

The options are mutually exclusive: setting include will override an exclude. Note that include must be an exhaustive list of files as otherwise necessary source files may not be included.

The publish field (optional)

The publish field can be used to prevent a package from being published to a repository by mistake.

# ...
publish = false

The workspace field (optional)

The workspace field can be used to configure the workspace that this package will be a member of. If not specified this will be inferred as the first Cargo.toml with [workspace] upwards in the filesystem.

# ...
workspace = "path/to/root"

For more information, see the documentation for the workspace table below.

Package metadata

There are a number of optional metadata fields also accepted under the [package] section:

# ...

# A short blurb about the package. This is not rendered in any format when
# uploaded to (aka this is not markdown).
description = "..."

# These URLs point to more information about the repository.
documentation = "..."
homepage = "..."
repository = "..."

# This points to a file in the repository (relative to this `Cargo.toml`). The
# contents of this file are stored and indexed in the registry.
readme = "..."

# This is a small list of keywords used to categorize and search for this
# package.
keywords = ["...", "..."]

# This is a string description of the license for this package. Currently
# will validate the license provided against a whitelist of known
# license identifiers from Multiple licenses can be
# separated with a `/`.
license = "..."

# If a project is using a nonstandard license, then this key may be specified in
# lieu of the above key and must point to a file relative to this manifest
# (similar to the readme key).
license-file = "..."

The registry will render the description, display the license, link to the three URLs and categorize by the keywords. These keys provide useful information to users of the registry and also influence the search ranking of a crate. It is highly discouraged to omit everything in a published crate.

The metadata table (optional)

Cargo by default will warn about unused keys in Cargo.toml to assist in detecting typos and such. The package.metadata table, however, is completely ignored by Cargo and will not be warned about. This section can be used for tools which would like to store project configuration in Cargo.toml. For example:

name = "..."
# ...

# Metadata used when generating an Android APK, for example.
package-name = "my-awesome-android-app"
assets = "path/to/static"

Dependency sections

See the specifying dependencies page for information on the [dependencies], [dev-dependencies], and target-specific [target.*.dependencies] sections.

The [profile.*] sections

Cargo supports custom configuration of how rustc is invoked through profiles at the top level. Any manifest may declare a profile, but only the top level project’s profiles are actually read. All dependencies’ profiles will be overridden. This is done so the top-level project has control over how its dependencies are compiled.

There are five currently supported profile names, all of which have the same configuration available to them. Listed below is the configuration available, along with the defaults for each profile.

# The development profile, used for `cargo build`.
opt-level = 0      # controls the `--opt-level` the compiler builds with
debug = true       # controls whether the compiler passes `-g`
rpath = false      # controls whether the compiler passes `-C rpath`
lto = false        # controls `-C lto` for binaries and staticlibs
debug-assertions = true # controls whether debug assertions are enabled
codegen-units = 1  # controls whether the compiler passes `-C codegen-units`
                   # `codegen-units` is ignored when `lto = true`
panic = 'unwind'   # panic strategy (`-C panic=...`), can also be 'abort'

# The release profile, used for `cargo build --release`.
opt-level = 3
debug = false
rpath = false
lto = false
debug-assertions = false
codegen-units = 1
panic = 'unwind'

# The testing profile, used for `cargo test`.
opt-level = 0
debug = true
rpath = false
lto = false
debug-assertions = true
codegen-units = 1
panic = 'unwind'

# The benchmarking profile, used for `cargo bench`.
opt-level = 3
debug = false
rpath = false
lto = false
debug-assertions = false
codegen-units = 1
panic = 'unwind'

# The documentation profile, used for `cargo doc`.
opt-level = 0
debug = true
rpath = false
lto = false
debug-assertions = true
codegen-units = 1
panic = 'unwind'

The [features] section

Cargo supports features to allow expression of:

A feature of a package is either an optional dependency, or a set of other features. The format for specifying features is:

name = "awesome"

# The default set of optional packages. Most people will want to use these
# packages, but they are strictly optional. Note that `session` is not a package
# but rather another feature listed in this manifest.
default = ["jquery", "uglifier", "session"]

# A feature with no dependencies is used mainly for conditional compilation,
# like `#[cfg(feature = "go-faster")]`.
go-faster = []

# The `secure-password` feature depends on the bcrypt package. This aliasing
# will allow people to talk about the feature in a higher-level way and allow
# this package to add more requirements to the feature in the future.
secure-password = ["bcrypt"]

# Features can be used to reexport features of other packages. The `session`
# feature of package `awesome` will ensure that the `session` feature of the
# package `cookie` is also enabled.
session = ["cookie/session"]

# These packages are mandatory and form the core of this package’s distribution.
cookie = "1.2.0"
oauth = "1.1.0"
route-recognizer = "=2.1.0"

# A list of all of the optional dependencies, some of which are included in the
# above `features`. They can be opted into by apps.
jquery = { version = "1.0.2", optional = true }
uglifier = { version = "1.5.3", optional = true }
bcrypt = { version = "*", optional = true }
civet = { version = "*", optional = true }

To use the package awesome:

version = "1.3.5"
default-features = false # do not include the default features, and optionally
                         # cherry-pick individual features
features = ["secure-password", "civet"]


The usage of features is subject to a few rules:

Note that it is explicitly allowed for features to not actually activate any optional dependencies. This allows packages to internally enable/disable features without requiring a new dependency.

Usage in end products

One major use-case for this feature is specifying optional features in end-products. For example, the Servo project may want to include optional features that people can enable or disable when they build it.

In that case, Servo will describe features in its Cargo.toml and they can be enabled using command-line flags:

$ cargo build --release --features "shumway pdf"

Default features could be excluded using --no-default-features.

Usage in packages

In most cases, the concept of optional dependency in a library is best expressed as a separate package that the top-level application depends on.

However, high-level packages, like Iron or Piston, may want the ability to curate a number of packages for easy installation. The current Cargo system allows them to curate a number of mandatory dependencies into a single package for easy installation.

In some cases, packages may want to provide additional curation for optional dependencies:

In almost all cases, it is an antipattern to use these features outside of high-level packages that are designed for curation. If a feature is optional, it can almost certainly be expressed as a separate package.

The [workspace] section

Projects can define a workspace which is a set of crates that will all share the same Cargo.lock and output directory. The [workspace] table can be defined as:


# Optional key, inferred if not present
members = ["path/to/member1", "path/to/member2"]

Workspaces were added to Cargo as part RFC 1525 and have a number of properties:

The root crate of a workspace, indicated by the presence of [workspace] in its manifest, is responsible for defining the entire workspace (listing all members). This can be done through the members key, and if it is omitted then members are implicitly included through all path dependencies. Note that members of the workspaces listed explicitly will also have their path dependencies included in the workspace.

The package.workspace manifest key (described above) is used in member crates to point at a workspace's root crate. If this key is omitted then it is inferred to be the first crate whose manifest contains [workspace] upwards in the filesystem.

A crate may either specify package.workspace or specify [workspace]. That is, a crate cannot both be a root crate in a workspace (contain [workspace]) and also be a member crate of another workspace (contain package.workspace).

Most of the time workspaces will not need to be dealt with as cargo new and cargo init will handle workspace configuration automatically.

The project layout

If your project is an executable, name the main source file src/ If it is a library, name the main source file src/

Cargo will also treat any files located in src/bin/*.rs as executables.

Your project can optionally contain folders named examples, tests, and benches, which Cargo will treat as containing example executable files, integration tests, and benchmarks respectively.

▾ src/           # directory containing source files         # the main entry point for libraries and packages        # the main entry point for projects producing executables
  ▾ bin/         # (optional) directory containing additional executables
▾ examples/      # (optional) examples
▾ tests/         # (optional) integration tests
▾ benches/       # (optional) benchmarks

To structure your code after you've created the files and folders for your project, you should remember to use Rust's module system, which you can read about in the book.


Files located under examples are example uses of the functionality provided by the library. When compiled, they are placed in the target/examples directory.

They must compile as executables (with a main() function) and load in the library by using extern crate <library-name>. They are compiled when you run your tests to protect them from bitrotting.

You can run individual examples with the command cargo run --example <example-name>.


When you run cargo test, Cargo will:

Integration tests

Each file in tests/*.rs is an integration test. When you run cargo test, Cargo will compile each of these files as a separate crate. The crate can link to your library by using extern crate <library-name>, like any other code that depends on it.

Cargo will not automatically compile files inside subdirectories of tests, but an integration test can import modules from these directories as usual. For example, if you want several integration tests to share some code, you can put the shared code in tests/common/ and then put mod common; in each of the test files.

Configuring a target

All of the [[bin]], [lib], [[bench]], [[test]], and [[example]] sections support similar configuration for specifying how a target should be built. The double-bracket sections like [[bin]] are array-of-table of TOML, which means you can write more than one [[bin]] section to make several executables in your crate.

The example below uses [lib], but it also applies to all other sections as well. All values listed are the defaults for that option unless otherwise specified.

# ...

# The name of a target is the name of the library that will be generated. This
# is defaulted to the name of the package or project, with any dashes replaced
# with underscores. (Rust `extern crate` declarations reference this name;
# therefore the value must be a valid Rust identifier to be usable.)
name = "foo"

# This field points at where the crate is located, relative to the `Cargo.toml`.
path = "src/"

# A flag for enabling unit tests for this target. This is used by `cargo test`.
test = true

# A flag for enabling documentation tests for this target. This is only relevant
# for libraries, it has no effect on other sections. This is used by
# `cargo test`.
doctest = true

# A flag for enabling benchmarks for this target. This is used by `cargo bench`.
bench = true

# A flag for enabling documentation of this target. This is used by `cargo doc`.
doc = true

# If the target is meant to be a compiler plugin, this field must be set to true
# for Cargo to correctly compile it and make it available for all dependencies.
plugin = false

# If the target is meant to be a "macros 1.1" procedural macro, this field must
# be set to true.
proc-macro = false

# If set to false, `cargo test` will omit the `--test` flag to rustc, which
# stops it from generating a test harness. This is useful when the binary being
# built manages the test runner itself.
harness = true

Building dynamic or static libraries

If your project produces a library, you can specify which kind of library to build by explicitly listing the library in your Cargo.toml:

# ...

name = "..."
crate-type = ["dylib"] # could be `staticlib` as well

The available options are dylib, rlib, staticlib, cdylib, and proc-macro. You should only use this option in a project. Cargo will always compile packages (dependencies) based on the requirements of the project that includes them.

You can read more about the different crate types in the Rust Reference Manual

The [replace] Section

This section of Cargo.toml can be used to override dependencies with other copies. The syntax is similar to the [dependencies] section:

"foo:0.1.0" = { git = '' }
"bar:1.0.2" = { path = 'my/local/bar' }

Each key in the [replace] table is a package id specification which allows arbitrarily choosing a node in the dependency graph to override. The value of each key is the same as the [dependencies] syntax for specifying dependencies, except that you can't specify features. Note that when a crate is overridden the copy it's overridden with must have both the same name and version, but it can come from a different source (e.g. git or a local path).

More information about overriding dependencies can be found in the overriding dependencies section of the documentation.