Skip to content

Releases: PStryder/MemoryGate

MemoryGate_Stanalone_v1.2.0

Choose a tag to compare

@PStryder PStryder released this 11 Jan 21:55
docs: add documentation set and refresh README

MemoryGate_Standalone_v1.1

Choose a tag to compare

@PStryder PStryder released this 10 Jan 23:55
v1.1 - Production Hardening Release

FEATURES:
- Structured error handling (ValidationIssue exception)
- Tool inventory monitoring with auto-rebind
- Strict pattern evidence validation
- Enhanced archive/rehydrate idempotency

FIXES:
- P0: Tool registry collapse detection and recovery
- P0: Archive/rehydrate accurate counts and status
- P1: Pattern evidence ID validation (strict mode)
- P1: Concept alias consistency across all tools
- P2: Structured validation errors (no stack traces)
- P2: Limits documented in user guide

DEPLOYMENT:
- Tested by Hexy (ChatGPT) and Kee (Claude Desktop)
- Multi-platform OAuth support validated
- Production stable at memorygate.fly.dev
- Ready for app store submission

Breaking Changes: None
Migration Required: Yes (8eaa0966fb0f - OAuth tables)

MemoryGate_Standalone

Choose a tag to compare

@PStryder PStryder released this 10 Jan 18:21

MemoryGate v1.0.0 - First Production Release
Forgetting That Remembers: Intelligent Memory Retention for AI Systems
This release introduces MemoryGate's forgetting mechanism - a production-ready system for managing AI memory through retention scoring, tier-based archival, and audit trails.
What's New
Retention Scoring System

Memories gain score when accessed (+0.4 per retrieval)
Score decays over time (-0.02 every 15 minutes)
Access patterns naturally surface important content

Tier-Based Archival

HOT tier: Actively used memories (score ≥ -1.0)
COLD tier: Archived summaries (score < -1.0)
Automatic migration based on access patterns

Soft Deletion with Audit Trails

Tombstone records preserve deletion metadata
Configurable purge thresholds (default: score < -2.0)
Full causality tracking for compliance

Technical Highlights

Migration 0002_cold_storage: Adds retention columns to all memory tables
Backward compatible: Existing data migrates seamlessly with default scores
Configurable parameters: Adjust bump/decay rates via environment variables
Production tested: Zero errors across 15 MCP tool integrations

Deployment
Production instance running at memorygate.fly.dev with PostgreSQL + pgvector backend. Schema revision 0002_cold_storage applied successfully.
Breaking Changes
None. Fully backward compatible with existing clients.

Retention Parameters:

SCORE_BUMP_ALPHA=0.4 (access boost)
SCORE_DECAY_BETA=0.02 (time decay)
SUMMARY_TRIGGER_SCORE=-1.0 (archive threshold)
PURGE_TRIGGER_SCORE=-2.0 (purge threshold)
RETENTION_TICK_SECONDS=900 (15 min cycles)