Metadata-Version: 2.4
Name: pactree
Version: 0.0.1
Summary: Deterministic context trees for AI agents and applications.
Project-URL: Homepage, https://github.com/earldennison/egregore-v2
Project-URL: Repository, https://github.com/earldennison/egregore-v2
Author: Earl Dennison Tan
License-Expression: MIT
Keywords: agent,context,pact,pactree,tree
Classifier: Development Status :: 2 - Pre-Alpha
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Programming Language :: Python :: 3.13
Classifier: Topic :: Software Development :: Libraries :: Python Modules
Requires-Python: >=3.11
Description-Content-Type: text/markdown

# Pactree

Pactree is a deterministic context-tree substrate for AI agents and
applications. It is designed to make context explicit: ordered, queryable,
branchable, compact on disk, and stable across Python and Rust runtimes.

The project centers on the `.pact` file format and a Rust engine for
context trees, operations, selectors, projections, traces, and rewind/fork
workflows. The Python package provides the ergonomic host API that applications
use to build, inspect, and persist those trees.

Core goals:

- deterministic tree layout for agent context
- compact content-addressed storage
- branch, rewind, and what-if workflows
- provider-facing projections without losing substrate structure
- a clean Python API backed by a canonical Rust implementation

This is an early `0.0.x` release while the standalone package is being split
out. The public namespace is available now; the Pactree engine binary and Python
bindings will land in the next extraction releases.
