DeepRail

The open standard for designing AI-native work, software, and organizations.

v0.16.8 — Public Release. The Japanese Standard projection is canonical; English normative modules are published as Preview.

DeepRail now connects the long-form Book to problem diagnostics, stable Standard modules, actor-neutral workflows, a stateful evidence-gated runtime, thin agent adapters, audit history, and reinvestment.

Start

Python 3.10 or newer is required. For runtime use, install the package in a virtual environment:

• Japanese Book — 45 chapters + ending
• Japanese 10-minute Quickstart
• Diagnostic
• English Quickstart
• Standard
• Workflows

python3 -m venv .venv
source .venv/bin/activate
pip install -e .
deeprail init --target /path/to/project --profile lightweight --communication en --artifacts en --adapter generic

Validate

For the full repository validation commands below, install the development extras first:

pip install -e ".[dev]"

./scripts/generate
./scripts/check
python3 -m unittest discover -s tests -v
python3 -m build
./scripts/build-dist

Release Status

• Public release: v0.16.8
• Canonical Standard projection: Japanese (ja), DR-M01〜DR-M25
• English Standard projection: Preview; normative module translation is planned for a later release
• Runtime communication and artifact output: Japanese and English
• Software license: MIT
• Documentation license: CC BY 4.0
• License scope: LICENSE-BOUNDARY.md
• Release history: CHANGELOG.md
• Support: GitHub Issues · Security

Authorship

Authorship and the public-content boundary are documented in AUTHORSHIP.md.