Skip to content

Dots Brain 0.4.4

Pre-release
Pre-release

Choose a tag to compare

@OnFiala OnFiala released this 18 Oct 14:43
· 23 commits to main since this release
0b43ebf

Dots Brain 0.4.4

Dots Brain is a self-hosted MCP memory service for explicitly saved notes. This is still a work in progress; the release is marked as a prerelease.

Changes

  • Make the default systemd example work with core dependencies; document explicit semantic dependencies and model preparation.
  • Verify fresh core-only wheel installations in Linux Python 3.11 and 3.12, including real stdio/HTTP MCP, template startup, migration and restore.
  • Bind release preparation to an annotated tag, its exact commit on main, current locked metadata and successful installed-wheel checks.
  • Reject incomplete indexing and full-text fallback in the semantic stress checks.
  • Preserve an existing empty or unknown restore disable barrier; allow only a valid retry of the same recovery.
  • Keep package, runtime, plugin, lockfile, README and changelog versions aligned. Schema v2 and dependency versions are unchanged.

This release also includes the 0.4.1–0.4.3 changes previously merged to main: ownership-aware cleanup, retrieval and documentation corrections, nginx request-buffering coverage, and the full-text replacement row-ID fix. See CHANGELOG.md for details.

Upgrade

Verify the supplied SHA256SUMS, install the selected artifact in a separate environment, and retain the previous executable and a validated private backup. Preserve an existing service's explicit semantic configuration. Upgrading does not repair an already damaged full-text index or remove recovery barriers. Follow docs/upgrading.md.

The source/plugin ZIP contains the reviewed repository files. Disabled capture and CORTEX modules remain excluded from the core wheel. Automatic conversation capture is not enabled.

Source commit: 0b43ebfaeb7d28fe1e4faf7f33e6b31924fb1afe.

Release publication and any installation are separate operations. Source and wheel checks do not establish recovery, client behavior or capacity for another deployment.