projectmem

projectmem enhances coding workflows by providing a local-first memory and judgment layer for development activities. Key features include: • Captures bugs, failed attempts, and fixes within your project • Warns at git commit to prevent repeating known dead-ends • Automatically classifies commits (revert as failed, fix as success) • Inject token-budgeted memory into coding sessions • Stores cross-project library insights and patterns This tool transforms stateless coding assistants into experienced engineers by providing a persistent context and learning mechanism. It operates entirely locally, is open source under the MIT license, and seamlessly integrates with popular coding tools such as Claude, Cursor, and Codex. By remembering past decisions, debugging efforts, and successful fixes, projectmem significantly reduces token consumption and eliminates the need for manual re-explanation of context. projectmem offers advanced features like pre-commit warnings that leverage its deep memory to alert developers before they re-introduce known issues. It automates the capture of project history, injecting relevant memory into current coding sessions, ensuring that past lessons inform present tasks. Additionally, a global memory layer stores library gotchas and reusable patterns across all projects, making new projects start with accumulated knowledge. Designed for individual developers and small engineering teams, projectmem is ideal for those seeking to maximize efficiency and consistency in their coding practices. It helps prevent repetitive errors, accelerate development cycles, and maintain a high standard of code quality without relying on cloud services or external APIs. Easily installable, it provides a transparent return on investment through provable debugging time and token savings.