Show HN: Engrim – A universal, local-first SQLite memory engine for AI CLIs

Engrim is a local-first SQLite memory engine designed for AI CLIs that allows developers to maintain persistent project context when switching between different AI models and environments without losing architectural decisions or project state. The tool uses hybrid search combining keyword matching and vector embeddings to consolidate episodic memory into ~1,000 tokens, reducing context reload costs by 99% compared to traditional approaches. In production testing on a 50,000-line codebase, engrim successfully maintained continuity across 105 sessions and multiple model switches without regressions or context loss.