Tree Ring Memory v0.12.0

Tree Ring Memory v0.12.0 remains protocol-preview and adds an explicit, controlled workflow-proof surface.

Highlights

• Controlled paired agent-workflow evaluation across no-memory, raw-memory, and Tree Ring retrieval arms.
• Retained per-trial workspaces, explicit model identity, and exact structured-output validation.
• Hardened proof output validation against symlink and hard-link substitution.
• Refreshed public certification evidence and release metadata.

Verification

• Tagged main CI passed.
• Tag-built Linux and macOS packages passed.
• Full local certification passed: install, CLI, performance, harness, recall-quality, and seven memory-quality scenarios.
• Package contents, checksums, and the macOS ARM64 binary version were verified after download.

Controlled workflow evidence

Codex CLI 0.144.0 with gpt-5.6-luna completed 9 retained trials:

• Tree Ring: 3/3
• Raw memory: 3/3
• No memory: 2/3

Tree Ring recorded one observed win over no memory and zero over raw memory. This is controlled fixture evidence, not a claim of universal superiority over raw context or all agent workflows.

Assets

• macOS ARM64 archive and SHA-256 checksum
• Linux x86_64 archive and SHA-256 checksum

Tree Ring Memory v0.11.0 Launch Preview

Tree Ring Memory v0.11.0 Launch Preview

Tree Ring Memory is a framework-agnostic, local-first memory lifecycle layer for
AI agents.

Launch page:
https://terminallylazy.github.io/Tree-Ring-Memory/

Feedback issue:
#26

Why This Release Exists

Agent memory should age, not pile up. Tree Ring Memory helps agents preserve
useful decisions, warnings, preferences, and evidence without turning memory
into raw transcript storage.

Fresh memory stays detailed. Older learning compresses into rings. Important
failures remain visible as scars. Durable truths become heartwood. Future ideas
stay as seeds.

What Ships

• Rust-native CLI and crates.
• Local SQLite/FTS storage.
• Explainable recall with ring, scope, confidence, and ranking signals.
• JSONL import/export.
• Audit for memory quality, privacy, and integrity.
• Deterministic consolidation without requiring an LLM.
• Rust-owned maintenance for expiry, secret redaction, and FTS repair.
• Source-linked evaluated outcomes through tree-ring evidence.
• DOX and Revolve sync adapters.
• Read-only agent-framework discovery.
• Terminal onboarding and a Ratatui operator console.

Install

Global user install:

curl -fsSL https://raw.githubusercontent.com/TerminallyLazy/Tree-Ring-Memory/main/install.sh | sh

Project-local install with initialization:

curl -fsSL https://raw.githubusercontent.com/TerminallyLazy/Tree-Ring-Memory/main/install.sh | sh -s -- --project --init

First Commands

tree-ring init
tree-ring remember "Use project-scoped recall before risky release changes." --event-type lesson --scope project
tree-ring recall "release changes"
tree-ring evidence "The eval passed after the fix." --outcome promoted --evidence-ref evals/run-042
tree-ring audit --audit-type sensitive
tree-ring tui

Release Artifact

This release includes a platform tarball built by scripts/package-release.sh
and a SHA-256 checksum file.

The installer can consume prebuilt archives through:

sh install.sh --archive-url <release-asset-url> --archive-sha256 <sha256>

Feedback Wanted

• Which agent frameworks should get first-class bridge support?
• What should explainable recall show by default?
• Where does the ring model feel too simple or too heavy?
• What privacy or forgetting control is missing?
• What makes the first ten minutes hard?