Skip to content

Releases: axtontc/OmniMem

v1.0.0 — Stable Release

Choose a tag to compare

@axtontc axtontc released this 18 Jul 07:50

OmniMem: Hybrid Graph-Vector Storage Engine for Infinite Memory

We are proud to announce the stable v1.0.0 release of OmniMem!

Key Updates

  • Permissive Apache 2.0 Licensing: Transitioned from PolyForm Noncommercial/MIT dual licensing to Apache 2.0 with copyright explicitly assigned to Axton Carroll.
  • Strict Static Typing: Annotated internal models (sanitized: Dict[str, Any] in memory router, self._queue and self._seen_txids in AsyncWAL) and resolved all mypy type errors.
  • None Safety Checks: Added non-null validation guards for database connections and embedding model instances in FastAPI /search and /search_graph endpoints.
  • Formatted and Cleaned Codebase: Cleaned up Ruff style issues (bare except blocks, colons, import groupings, unused variables) and ignore E501 limit on long SQL/Cypher database queries.
  • Unified CI/CD Workflow: Replaced legacy setup jobs with a workflow running lint checks and multi-python version test suites using setup-uv runner caching.
  • Premium Landing Page: Updated README with badges, design flow chart, CLI reference guides, subsystems grid, API reference table, and related projects cross-linking.

Full Changelog: https://github.com/axtontc/OmniMem/commits/v1.0.0