Anchored

Anchored sets up a local-first memory database on your machine that speaks the Model Context Protocol (MCP). It allows AI coding assistants—such as Claude Code, Cursor, Gemini CLI, VS Code Copilot, and Codex—to read and write shared memories, bridging the context gap between tools. This helps developers maintain a continuous workflow when switching between assistants.

Key strengths: it is a single, self-contained binary with zero runtime dependencies; no account or cloud connection required, ensuring privacy and simplicity. For power users who juggle multiple AI coding tools, it can significantly reduce redundant explanations and context loss.

Downsides: the project is extremely young (6 GitHub stars, minimal community), so its feature depth, documentation, and long-term maintenance are uncertain. Setup may require manual configuration and only works with MCP-compatible tools. If you only use one coding assistant, the added complexity may not be justified.

Anchored is open-source and free. It suits developers who routinely use several AI coding tools and want persistent, cross-tool memory.