OpenLore

•
367
•

Provides persistent architectural memory and structural cognition for AI coding agents by transforming codebases into queryable knowledge graphs.

AI coding agents often struggle with context decay, repeatedly re-reading code, forgetting architectural decisions, and incurring high token costs for orientation. OpenLore addresses these challenges by transforming any codebase into a navigable knowledge graph, providing persistent architectural memory and structural cognition across agent sessions. It integrates static analysis to build call graphs and detect clusters, facilitates LLM-generated living specifications via OpenSpec for semantic intent and automated drift detection, and exposes a suite of 45 graph-native MCP tools for efficient agent interaction, drastically reducing token costs and ensuring architectural assumptions remain fresh.

Key Features

01Static analysis for call graphs, code clusters, and architectural digests
02Automated drift detection between code and specifications with pre-commit hooks
03Auto-configuration for popular AI coding agents to integrate `orient()` calls
04137 GitHub stars
05LLM-generated living specifications and architectural decision records (ADRs)
06Persistent architectural memory and token-efficient orientation for AI agents

Use Cases

01Preventing context decay and architectural amnesia in AI coding agent workflows
02Optimizing token expenditure for AI agents by providing targeted, queryable architectural context
03Automating the detection of spec-code drift and ensuring living documentation consistency
04Managing architectural decisions and gating code changes based on review