Get started

Standard is an alpha Python library published on PyPI as standard; its Python package and command are named std. Install the public package or install the source checkout as shown below.

Install from PyPI

python -m pip install standard
python -m pip install 'standard[fastapi]'
python -m pip install 'standard[aws]'
python -m pip install 'standard[http]'
python -m pip install 'standard[cache]'
python -m pip install 'standard[rdbms]'
python -m pip install 'standard[fastapi,aws,http,cache,rdbms]'

The current coordinated release is 0.2.4. The core distribution does not install optional integration packages unless you select an extra. You can also install one provider directly with python -m pip install standard-fastapi.

Install from this checkout

Run these commands from the repository root. Create and activate a virtual environment, then install the core package in editable mode:

python3 -m venv .venv
. .venv/bin/activate
python -m pip install -e .

On Windows, activate the environment with .venv\Scripts\activate instead. Install a local integration project when you need it:

These commands install this checkout in editable mode. Use them when changing Standard itself; application projects should normally install the published distribution and the extras they use.

Load a configuration file

Create config/settings.yaml:

database:
  host: localhost

Read the nested value:

from std.config import ConfigManager

config = ConfigManager.from_file("config/settings.yaml")
host = config.snapshot.get_path("database.host", "localhost")
print(host)

from_file accepts one YAML or TOML file, or a directory containing those files. Missing sources raise MissingConfigSourceError by default; pass optional=True only when a missing source is expected. See the configuration reference for merge and interpolation rules.

Use Pydantic Settings

Create config/app.yaml:

host: localhost
port: 8080

Then define a Settings class:

from std.config import YamlSettings


class Settings(YamlSettings):
    std_config_file = "config/app.yaml"
    host: str
    port: int = 8080

Define this model in core/settings.py, then construct it once from main.py:

from app.core.settings import Settings


def main() -> None:
    settings = Settings()
    # Build the application and inject its validated settings here.


if __name__ == "__main__":
    main()

main.py is the only place that calls Settings(). YamlSettings reads settings.yaml by default; std_config_file selects another file. TomlSettings works the same way and defaults to settings.toml. Both convenience bases use this source priority: constructor values, config file, process environment, dotenv, then file secrets. The first source has the highest priority. A custom settings_customise_sources hook can change that order.

Configure application logging

std.logging uses environment-aware defaults and lets explicit settings override the level and format independently:

from std.logging import LoggingConfig, configure_logging

configure_logging(
    LoggingConfig(),
    environment="production",
    service_name="billing-api",
)

Development uses DEBUG and readable console output. Tests use WARNING and console output. Staging and production use INFO and JSON lines. The default stream is stdout; JSON records include timestamp, severity, logger, message, environment, optional service/source fields, exception details, and structured extra fields. See the logging guide for YAML-backed settings and custom formats.

Agent commands

std guide
std guide --format json
std skill show
std skill install

The skill install command writes .agents/skills/standard-idiomatic-python/SKILL.md by default. The directory name matches the skill's required Agent Skills metadata. The command refuses to replace an existing file; pass --force to replace it or give another destination path. See the CLI reference.

Resolved in 0.2.4: the 0.2.3 package's installed skill contained one stale sentence that described std.service as unreleased, although std guide listed it as part of core. The 0.2.4 package corrects that sentence and its published skill was verified after installation. See testing for the release-artifact evidence.

Current status

The repository's publishing guide lists all six distributions, their import paths, and the shared tag release process. The testing guide distinguishes source-tree checks from consumer and release-artifact checks.

Read the docs

• Known means supported by source code, project configuration, or a linked source.
• Proposal means a design option that has not been agreed.
• Open means an external setup task or project decision remains.
• Measured means the result has raw benchmark data and a reproducible setup.

Continue with the configuration reference, API reference, testing and benchmarks, or contributing guide.