Skip to main content

aircore / airpy / airlang / aircli

A provider-agnostic AI execution runtime, split into four independently installable packages with a strict, one-directional dependency chain:

aircore   <-  airpy   <-  airlang
   ^           ^             ^
   └───────────┴─────────────┘
              aircli
  • aircore -- the execution runtime: scheduler, capabilities, policy, journal, consensus, memory, approval, checkpoint/resume, sandboxed execution, cross-step data flow. No model, prompt, or provider concept exists anywhere in this package, by design. Zero dependencies.
  • airpy -- the provider-aware Python SDK on top of aircore: ModelAgent, real provider adapters (LiteLLMProvider, native OpenAIProvider), consensus strategies, structured output, an MCP tool registry, streaming, long-running sessions, PromptTemplate. Imports aircore; aircore never imports airpy.
  • airlang -- AirLang, a declarative workflow language: lexer, parser, IR, and an executor that compiles a parsed .airlang file to real airpy/aircore calls. Imports airpy; airpy never imports airlang. See airlang-spec-v1.md.
  • aircli -- the ai command: ai run/ai trace for both .py scripts and .airlang files, with --json/--html trace output. Depends on all three of the above.

Install

airpy's and aircli's PyPI distribution names differ from their import names -- airpy/aircli were both too close to existing PyPI project names and got rejected, so they're published as airpyy / airclii. You still import airpy and run the ai command exactly as before; only the pip install name changes.

pip install aircore           # just the runtime
pip install airpyy            # runtime + SDK           (import as: airpy)
pip install airlang           # runtime + SDK + language frontend
pip install airclii           # everything, plus the `ai` command

Optional extras on airpy: litellm, pydantic, mcp, openai -- each is a lazy import inside the one module that needs it, so none of them are required just to install airpy itself. (pip install "airpyy[litellm]" etc.)

For local development on this repo (all four packages, editable, pip install -e . style), see the root pyproject.toml instead of the per-package ones under packaging/ -- those are used only by the release workflow (.github/workflows/publish.yml) to build the four separate PyPI distributions.

Docs

  • architecture-spec-v1.md -- the runtime's original design spec.
  • airlang-spec-v1.md -- AirLang's design spec.
  • cross-step-data-flow.md -- the bindings/let design writeup.
  • real-project-readiness-roadmap.md -- known gaps and what's planned.

Metadata

Release files for airlang 0.4.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for airlang 0.4.1
File Size Uploaded Provenance Actions
airlang-0.4.1.tar.gz 18.9 kB With provenance Details for airlang-0.4.1.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for airlang 0.4.1
File Interpreter ABI Platform Actions
airlang-0.4.1-py3-none-any.whl Python 3 none any Details for airlang-0.4.1-py3-none-any.whl

Total release size: 39.0 kB

Release files / airlang-0.4.1.tar.gz

Download URL airlang-0.4.1.tar.gz
Size 18.9 kB
Tags Source
SHA-256 checksum
How to use checksums
da0137dee9f69173332e80e24cae6cf34e22a071a0610871319938b981969171
BLAKE2b-256 checksum
How to use checksums
72869b262f45fa927905c8151de7054d11a1880aef82c6aa35d0b9347e251f3e
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 19, 2026.

Transparency log

Release files / airlang-0.4.1-py3-none-any.whl

Download URL airlang-0.4.1-py3-none-any.whl
Size 20.1 kB
Tags Python 3
SHA-256 checksum
How to use checksums
82d4a8575684744164af3cead6fe07478c30a0cee02ddaad73e69cefd02477d4
BLAKE2b-256 checksum
How to use checksums
a49ced6c15baddccc6db4e5138ec83e2ddc07f843801724d090017a07aeb7ee0
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 19, 2026.

Transparency log

Security for airlang 0.4.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

Full provenance

Both files in this release have been uploaded with provenance metadata.

Select from 2 files

Attestation for airlang-0.4.1.tar.gz

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 19, 2026.

Transparency log

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.4.1 This release

2 release files Full provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page