Engraphis open-sources a local-first bitemporal memory tool for AI coding agents

Engraphis has launched a local-first memory tool for AI coding agents, powered by SQLite for fully offline operation and integrated with diverse agents via the Model Context Protocol (MCP). The system combines vector, lexical, graph, and code-aware hybrid retrieval, fully preserving the history of fact changes over time instead of overwriting them, and actively declines to answer when evidence is insufficient. The project's core is open-sourced under the Apache-2.0 license and includes a local visual dashboard.