Configuring projects

Python version requirement

Projects may declare the Python versions supported by the project in the project.requires-python field of the pyproject.toml.

It is recommended to set a requires-python value:

pyproject.toml

[project]
name = "example"
version = "0.1.0"
requires-python = ">=3.12"

The Python version requirement determines the Python syntax that is allowed in the project and affects selection of dependency versions (they must support the same Python version range).

Entry points

Entry points are the official term for an installed package to advertise interfaces. These include:

• Command line interfaces
• Graphical user interfaces
• Plugin entry points

Important

Using the entry point tables requires a build system to be defined.

Command-line interfaces

Projects may define command line interfaces (CLIs) for the project in the [project.scripts] table of the pyproject.toml.

For example, to declare a command called hello that invokes the hello function in the example module:

pyproject.toml

[project.scripts]
hello = "example:hello"

Then, the command can be run from a console:

$ uv run hello

Graphical user interfaces

Projects may define graphical user interfaces (GUIs) for the project in the [project.gui-scripts] table of the pyproject.toml.

Important

These are only different from command-line interfaces on Windows, where they are wrapped by a GUI executable so they can be started without a console. On other platforms, they behave the same.

For example, to declare a command called hello that invokes the app function in the example module:

pyproject.toml

[project.gui-scripts]
hello = "example:app"

Plugin entry points

Projects may define entry points for plugin discovery in the [project.entry-points] table of the pyproject.toml.

For example, to register the example-plugin-a package as a plugin for example:

pyproject.toml

[project.entry-points.'example.plugins']
a = "example_plugin_a"

Then, in example, plugins would be loaded with:

example/__init__.py

from importlib.metadata import entry_points

for plugin in entry_points(group='example.plugins'):
    plugin.load()

Note

The group key can be an arbitrary value, it does not need to include the package name or "plugins". However, it is recommended to namespace the key by the package name to avoid collisions with other packages.

Build systems

A build system determines how the project should be packaged and installed. Projects may declare and configure a build system in the [build-system] table of the pyproject.toml.

uv uses the presence of a build system to determine if a project contains a package that should be installed in the project virtual environment. If a build system is not defined, uv will not attempt to build or install the project itself, just its dependencies. If a build system is defined, uv will build and install the project into the project environment.

The --build-backend option can be provided to uv init to create a packaged project with an appropriate layout. The --package option can be provided to uv init to create a packaged project with the default build system.

Note

While uv will not build and install the current project without a build system definition, the presence of a [build-system] table is not required in other packages. For legacy reasons, if a build system is not defined, then setuptools.build_meta:__legacy__ is used to build the package. Packages you depend on may not explicitly declare their build system but are still installable. Similarly, if you add a dependency on a local project or install it with uv pip, uv will attempt to build and install it regardless of the presence of a [build-system] table.

Build system options

Build systems are used to power the following features:

• Including or excluding files from distributions
• Editable installation behavior
• Dynamic project metadata
• Compilation of native code
• Vendoring shared libraries

To configure these features, refer to the documentation of your chosen build system.

Project packaging

As discussed in build systems, a Python project must be built to be installed. This process is generally referred to as "packaging".

You probably need a package if you want to:

• Add commands to the project
• Distribute the project to others
• Use a src and test layout
• Write a library

You probably do not need a package if you are:

• Writing scripts
• Building a simple application
• Using a flat layout

While uv usually uses the declaration of a build system to determine if a project should be packaged, uv also allows overriding this behavior with the tool.uv.package setting.

Setting tool.uv.package = true will force a project to be built and installed into the project environment. If no build system is defined, uv will use the setuptools legacy backend.

Setting tool.uv.package = false will force a project package not to be built and installed into the project environment. uv will ignore a declared build system when interacting with the project; however, uv will still respect explicit attempts to build the project such as invoking uv build.

Project environment path

The UV_PROJECT_ENVIRONMENT environment variable can be used to configure the project virtual environment path (.venv by default).

If a relative path is provided, it will be resolved relative to the workspace root. If an absolute path is provided, it will be used as-is, i.e., a child directory will not be created for the environment. If an environment is not present at the provided path, uv will create it.

This option can be used to write to the system Python environment, though it is not recommended. uv sync will remove extraneous packages from the environment by default and, as such, may leave the system in a broken state.

To target the system environment, set UV_PROJECT_ENVIRONMENT to the prefix of the Python installation. For example, on Debian-based systems, this is usually /usr/local:

$ python -c "import sysconfig; print(sysconfig.get_config_var('prefix'))"
/usr/local

To target this environment, you'd export UV_PROJECT_ENVIRONMENT=/usr/local.

Important

If an absolute path is provided and the setting is used across multiple projects, the environment will be overwritten by invocations in each project. This setting is only recommended for use for a single project in CI or Docker images.

Note

By default, uv does not read the VIRTUAL_ENV environment variable during project operations. A warning will be displayed if VIRTUAL_ENV is set to a different path than the project's environment. The --active flag can be used to opt-in to respecting VIRTUAL_ENV. The --no-active flag can be used to silence the warning.

Build isolation

By default, uv builds all packages in isolated virtual environments alongside their declared build dependencies, as per PEP 517.

Some packages are incompatible with this approach to build isolation, be it intentionally or unintentionally.

For example, packages like flash-attn and deepspeed need to build against the same version of PyTorch that is installed in the project environment; by building them in an isolated environment, they may inadvertently build against a different version of PyTorch, leading to runtime errors.

In other cases, packages may accidentally omit necessary dependencies in their declared build dependency list. For example, cchardet requires cython to be installed in the project environment prior to installing cchardet, but does not declare it as a build dependency.

To address these issues, uv supports two separate approaches to modifying the build isolation behavior:

1. Augmenting the list of build dependencies: This allows you to install a package in an isolated environment, but with additional build dependencies that are not declared by the package itself via the extra-build-dependencies setting. For packages like flash-attn, you can even enforce that those build dependencies (like torch) match the version of the package that is or will be installed in the project environment.

2. Disabling build isolation for specific packages: This allows you to install a package without building it in an isolated environment.

When possible, we recommend augmenting the build dependencies rather than disabling build isolation entirely, as the latter approach requires that the build dependencies are installed in the project environment prior to installing the package itself, which can lead to more complex installation steps, the inclusion of extraneous packages in the project environment, and difficulty in reproducing the project environment in other contexts.

Augmenting build dependencies

To augment the list of build dependencies for a specific package, add it to the extra-build-dependencies list in your pyproject.toml.

For example, to build cchardet with cython as an additional build dependency, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["cchardet"]

[tool.uv.extra-build-dependencies]
cchardet = ["cython"]

To ensure that a build dependency matches the version of the package that is or will be installed in the project environment, set match-runtime = true in the extra-build-dependencies table. For example, to build deepspeed with torch as an additional build dependency, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["deepspeed", "torch"]

[tool.uv.extra-build-dependencies]
deepspeed = [{ requirement = "torch", match-runtime = true }]

This will ensure that deepspeed is built with the same version of torch that is installed in the project environment.

Similarly, to build flash-attn with torch as an additional build dependency, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["flash-attn", "torch"]

[tool.uv.extra-build-dependencies]
flash-attn = [{ requirement = "torch", match-runtime = true }]

[tool.uv.extra-build-variables]
flash-attn = { FLASH_ATTENTION_SKIP_CUDA_BUILD = "TRUE" }

Note

The FLASH_ATTENTION_SKIP_CUDA_BUILD environment variable ensures that flash-attn is installed from a compatible, pre-built wheel, rather than attempting to build it from source, which requires access to the CUDA development toolkit. If the CUDA toolkit is not available, the environment variable can be omitted, and flash-attn will be installed from a pre-built wheel if one is available for the current platform, Python version, and PyTorch version.

Similarly, deep_gemm follows the same pattern:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["deep_gemm", "torch"]

[tool.uv.sources]
deep_gemm = { git = "https://github.com/deepseek-ai/DeepGEMM" }

[tool.uv.extra-build-dependencies]
deep_gemm = [{ requirement = "torch", match-runtime = true }]

The use of extra-build-dependencies and extra-build-variables are tracked in the uv cache, such that changes to these settings will trigger a reinstall and rebuild of the affected packages. For example, in the case of flash-attn, upgrading the version of torch used in your project would subsequently trigger a rebuild of flash-attn with the new version of torch.

Dynamic metadata

The use of match-runtime = true is only available for packages like flash-attn that declare static metadata. If static metadata is unavailable, uv is required to build the package during the dependency resolution phase; as such, uv cannot determine the version of the build dependency that would ultimately be installed in the project environment.

In other words, if flash-attn did not declare static metadata, uv would not be able to determine the version of torch that would be installed in the project environment, since it would need to build flash-attn prior to resolving the torch version.

As a concrete example, axolotl is a popular package that requires augmented build dependencies, but does not declare static metadata, as the package's dependencies vary based on the version of torch that is installed in the project environment. In this case, users should instead specify the exact version of torch that they intend to use in their project, and then augment the build dependencies with that version.

For example, to build axolotl against torch==2.6.0, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["axolotl[deepspeed, flash-attn]", "torch==2.6.0"]

[tool.uv.extra-build-dependencies]
axolotl = ["torch==2.6.0"]
deepspeed = ["torch==2.6.0"]
flash-attn = ["torch==2.6.0"]

Similarly, older versions of flash-attn did not declare static metadata, and thus would not have supported match-runtime = true out of the box. Unlike axolotl, though, flash-attn did not vary its dependencies based on dynamic properties of the build environment. As such, users could instead provide the flash-attn metadata upfront via the dependency-metadata setting, thereby forgoing the need to build the package during the dependency resolution phase. For example, to provide the flash-attn metadata upfront:

pyproject.toml

[[tool.uv.dependency-metadata]]
name = "flash-attn"
version = "2.6.3"
requires-dist = ["torch", "einops"]

Tip

To determine the package metadata for a package like flash-attn, navigate to the appropriate Git repository, or look it up on PyPI and download the package's source distribution. The package requirements can typically be found in the setup.py or setup.cfg file.

(If the package includes a built distribution, you can unzip it to find the METADATA file; however, the presence of a built distribution would negate the need to provide the metadata upfront, since it would already be available to uv.)

The version field in tool.uv.dependency-metadata is optional for registry-based dependencies (when omitted, uv will assume the metadata applies to all versions of the package), but required for direct URL dependencies (like Git dependencies).

Disabling build isolation

Installing packages without build isolation requires that the package's build dependencies are installed in the project environment prior to building the package itself.

For example, historically, to install cchardet without build isolation, you would first need to install the cython and setuptools packages in the project environment, followed by a separate invocation to install cchardet without build isolation:

$ uv venv
$ uv pip install cython setuptools
$ uv pip install cchardet --no-build-isolation

uv simplifies this process by allowing you to specify packages that should not be built in isolation via the no-build-isolation-package setting in your pyproject.toml and the --no-build-isolation-package flag in the command line. Further, when a package is marked for disabling build isolation, uv will perform a two-phase install, first installing any packages that support build isolation, followed by those that do not. As a result, if a project's build dependencies are included as project dependencies, uv will automatically install them before installing the package that requires build isolation to be disabled.

For example, to install cchardet without build isolation, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["cchardet", "cython", "setuptools"]

[tool.uv]
no-build-isolation-package = ["cchardet"]

When running uv sync, uv will first install cython and setuptools in the project environment, followed by cchardet (without build isolation):

$ uv sync --extra build
 + cchardet==2.1.7
 + cython==3.1.3
 + setuptools==80.9.0

Similarly, to install flash-attn without build isolation, include the following in your pyproject.toml:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["flash-attn", "torch"]

[tool.uv]
no-build-isolation-package = ["flash-attn"]

When running uv sync, uv will first install torch in the project environment, followed by flash-attn (without build isolation). As torch is both a project dependency and a build dependency, the version of torch is guaranteed to be consistent between the build and runtime environments.

A downside of the above approach is that it requires the build dependencies to be installed in the project environment, which is appropriate for flash-attn (which requires torch both at build-time and runtime), but not for cchardet (which only requires cython at build-time).

To avoid including build dependencies in the project environment, uv supports a two-step installation process that allows you to separate the build dependencies from the packages that require them.

For example, the build dependencies for cchardet can be isolated to an optional build group, as in:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["cchardet"]

[project.optional-dependencies]
build = ["setuptools", "cython"]

[tool.uv]
no-build-isolation-package = ["cchardet"]

Given the above, a user would first sync with the build optional group, and then without it to remove the build dependencies:

$ uv sync --extra build
 + cchardet==2.1.7
 + cython==3.1.3
 + setuptools==80.9.0
$ uv sync
 - cython==3.1.3
 - setuptools==80.9.0

Some packages, like cchardet, only require build dependencies for the installation phase of uv sync. Others require their build dependencies to be present even just to resolve the project's dependencies during the resolution phase.

In such cases, the build dependencies can be installed prior to running any uv lock or uv sync commands, using the lower lower-level uv pip API. For example, given:

pyproject.toml

[project]
name = "project"
version = "0.1.0"
description = "..."
readme = "README.md"
requires-python = ">=3.12"
dependencies = ["flash-attn"]

[tool.uv]
no-build-isolation-package = ["flash-attn"]

You could run the following sequence of commands to sync flash-attn:

$ uv venv
$ uv pip install torch setuptools
$ uv sync

Alternatively, users can instead provide the flash-attn metadata upfront via the dependency-metadata setting, thereby forgoing the need to build the package during the dependency resolution phase. For example, to provide the flash-attn metadata upfront:

pyproject.toml

[[tool.uv.dependency-metadata]]
name = "flash-attn"
version = "2.6.3"
requires-dist = ["torch", "einops"]

Editable mode

By default, the project will be installed in editable mode, such that changes to the source code are immediately reflected in the environment. uv sync and uv run both accept a --no-editable flag, which instructs uv to install the project in non-editable mode. --no-editable is intended for deployment use-cases, such as building a Docker container, in which the project should be included in the deployed environment without a dependency on the originating source code.

Conflicting dependencies

uv resolves all project dependencies together, including optional dependencies ("extras") and dependency groups. If dependencies declared in one section are not compatible with those in another section, uv will fail to resolve the requirements of the project with an error.

uv supports explicit declaration of conflicting dependency groups. For example, to declare that the optional-dependency groups extra1 and extra2 are incompatible:

pyproject.toml

[tool.uv]
conflicts = [
    [
      { extra = "extra1" },
      { extra = "extra2" },
    ],
]

Or, to declare the development dependency groups group1 and group2 incompatible:

pyproject.toml

[tool.uv]
conflicts = [
    [
      { group = "group1" },
      { group = "group2" },
    ],
]

See the resolution documentation for more.

Limited resolution environments

If your project supports a more limited set of platforms or Python versions, you can constrain the set of solved platforms via the environments setting, which accepts a list of PEP 508 environment markers. For example, to constrain the lockfile to macOS and Linux, and exclude Windows:

pyproject.toml

[tool.uv]
environments = [
    "sys_platform == 'darwin'",
    "sys_platform == 'linux'",
]

See the resolution documentation for more.

Required environments

If your project must support a specific platform or Python version, you can mark that platform as required via the required-environments setting. For example, to require that the project supports Intel macOS:

pyproject.toml

[tool.uv]
required-environments = [
    "sys_platform == 'darwin' and platform_machine == 'x86_64'",
]

The required-environments setting is only relevant for packages that do not publish a source distribution (like PyTorch), as such packages can only be installed on environments covered by the set of pre-built binary distributions (wheels) published by that package.

See the resolution documentation for more.