Mercury Cortex: A Local-First Knowledge Engine for AI Coding Assistants

Overview

Mercury Cortex is a tool designed to solve a common problem for developers using AI coding assistants. Tools like Claude Code or Copilot are powerful, but they often forget what you told them in previous conversations or struggle to understand the full history of a large project. Mercury Cortex fixes this by acting as a local memory engine. It builds a structured knowledge graph from your actual project files. This allows your AI assistant to remember file relationships, code patterns, and project architecture over time. The entire system runs on your own machine, which means your code never leaves your device. It connects to any AI assistant that supports the Model Context Protocol, creating a unified context for all your development work.

Benefits

Mercury Cortex offers several key advantages for developers who want to work more efficiently with AI tools.

• Persistent Project Memory:You can register a project once, and the engine will remember it forever. When you return to a project after weeks, your AI partner will recall file dependencies, code conventions, and specific implementations. This eliminates the need to re-explain your project structure every time you start a new chat.

• Cross-Project Knowledge:Unlike standard local search, Mercury Cortex maintains a single knowledge graph across all your registered repositories. This allows your AI to search through every indexed project to find existing code before writing new lines. You can reuse battle-tested utilities and components from other projects instead of rewriting them from scratch.

• Semantic Search:The tool allows you to search by what the code does rather than just filenames. You can look for functions by their purpose, features, or tags. The results point back to the original source location, so the AI can read the full code before adapting it to your current needs.

• Local and Secure:Because the engine runs entirely on your machine, your intellectual property stays safe. There is no risk of sensitive code being sent to a cloud server for processing.

Use Cases

Mercury Cortex is ideal for developers who work with large codebases or multiple projects simultaneously.

• Large-Scale Development:If you are working on a complex application with many files, Mercury Cortex helps the AI understand the entire architecture. This leads to more accurate suggestions and fewer errors when the AI generates new code.

• Code Reuse and Refactoring:Developers can use the tool to find similar implementations across different projects. This is useful for standardizing code patterns or refactoring legacy codebases by pulling in proven solutions from other repositories.

• Onboarding and Continuity:When a developer returns to a project after a long break, the AI can quickly catch up on the latest changes. This reduces the time spent reorienting and allows the developer to focus on new features.

• Integration with Existing Tools:The tool works with popular AI coding assistants like OpenCode, Claude Code, and Google Antigravity. Developers can add this memory layer to their existing workflow without changing how they write code.

Pricing

Mercury Cortex is an open-source project released under the Apache-2.0 license. It is free to download and use. There are no subscription fees or hidden costs. Users can install it via prebuilt binaries for Linux, macOS, and Windows, or build it from source using Rust.

Vibes

As an open-source tool, Mercury Cortex does not have public reviews or testimonials in the traditional sense. However, the project documentation highlights its potential to transform how developers interact with AI. The creators emphasize that this tool bridges the gap between powerful AI models and the practical needs of real-world software development. The community focus is on building a platform that evolves from a personal tool into an organization-wide knowledge system. Early feedback from the development community suggests that the ability to maintain context across long-term projects is a significant improvement over current AI coding assistants.

Additional Information

Mercury Cortex is built on the SurrealDB database for local storage and uses the Model Context Protocol to connect with AI tools. The project is currently in active development with a clear long-term vision. Future plans include features for team collaboration, where multiple developers can share a common knowledge graph. The team also aims to enable multi-agent collaboration, allowing different AI assistants to coordinate through the knowledge graph. The project is maintained by the Mercury Cortex Contributors and is available on GitHub.