AICL v2.1.0 — Architecture Compilation Language
AICL v2.1.0 — Architecture Compilation Language
Auditable compilation with Proof of Origin. If the compiler cannot explain why, it does not generate.
What's new in this release
This release is a comprehensive professionalisation of the AICL repository. The repo has been restructured into a clean python/ + editor/ workspace (still a single repo), all 11 upgrade suggestions from the audit have been implemented, and the project is now ready for the upcoming CogNet integration.
Repository structure
python/— the AICL language (compiler, CLI, TUI, examples, tests, tools)editor/— the Next.js 16 web IDE (optional)docs/— cross-cutting documentation, papers, screenshots, asciinema demo.github/workflows/— CI on Python 3.10, 3.11, 3.12 + editor build
Headline features
- 9-stage compiler producing auditable Python / Rust / JavaScript / Go
- 100% audit coverage — every generated artifact carries provenance
- Cryptographic Proof of Origin (Ed25519-signed) for every compilation
- Self-healing runtime — declared Risks bound to declared Recoveries at compile time
- 27 keywords across 10 language levels — formal grammar in
python/spec/grammar.md - Autonomous compilation loop (
aicl evolve) — self-writing, self-validating - Optional Next.js 16 web editor with live compilation, architecture tree visualisation, AI-assisted spec authoring
Quality engineering
- 156-test pytest suite (151 functional + 5 Hypothesis property-based)
- ruff + black + mypy configured in
python/pyproject.toml - pre-commit hooks for Python and TypeScript (
.pre-commit-config.yaml) - CI on GitHub Actions for Python 3.10/3.11/3.12 + editor build
- PEP 561
py.typedmarker for type-checker compatibility - Formal bridge protocol documented (
python/docs/bridge_protocol.md)
Documentation
- Concise 259-line README (down from 1044)
- ARCHITECTURE.md explaining the monorepo layout and data flow
- CHANGELOG.md in Keep-a-Changelog format (history back to v1.0.0)
- CONTRIBUTING.md with tooling section (bun vs npm, pre-commit, CI)
- SECURITY.md vulnerability disclosure policy
- CITATION.cff for academic citation
- CogNet integration plan (
python/docs/cognet_integration_plan.md) - arXiv paper, position paper, user/CLI/editor manuals (all in
docs/)
CogNet integration (ready)
A placeholder python/src/aicl/cognet/ subpackage is reserved so the import path is stable across the integration rollout. The full 4-phase plan is in python/docs/cognet_integration_plan.md. The six CogNet integration specs live in python/examples/showcase/cognet/.
Install
git clone https://github.com/AFKmoney/AICL.git
cd AICL
make install-python # or: cd python && pip install -e ".[tui,dev]"Quick start
aicl compile python/examples/showcase/01_blue_square.aicl
aicl tree python/examples/showcase/01_blue_square.aicl
aicl verify python/examples/showcase/01_blue_square.aiclSee the README for the full feature list and the CHANGELOG for full history.
Author: Philippe-Antoine Robert (@AFKmoney)
License: MIT
Tag: v2.1.0
Commit: 5e7c200