Cortex
VerifiedAutomatic memory system that preserves project context for AI coding assistants across sessions.
What is Cortex?
Cortex is an open-source VS Code extension that builds persistent memory for AI coding assistants. It observes interactions to record key elements such as architectural choices, bug patterns, and file modifications, then organizes them into working, episodic, and semantic layers for later retrieval.
The system runs in the background during sessions, performing periodic extractions and injecting summarized context before the next interaction begins. It supports optional LLM providers for deeper analysis while offering local pattern matching as a fallback.
It targets developers who frequently restart AI-assisted coding sessions and need to avoid repeating project explanations or losing prior context.
Capabilities
What you can build with Cortex
Seamless Session Resumption
Start a new coding session with full awareness of past decisions and open issues already loaded into the AI's context.
Decision Tracking
Automatically log architectural choices with reasons, alternatives, and affected files for future reference.
Bug Pattern Avoidance
Record root causes from previous fixes so the AI avoids repeating similar mistakes in ongoing work.
Install Cortex
ext install cortex-dev.cortex-memoryext install cortex-dev.cortex-memory- 1Install the Cortex Memory extension from the VS Code marketplace.
- 2Optionally configure an API key for enhanced extraction via the command palette.
- 3Open your project and begin working with an AI assistant like Claude or Cursor.
- 4Monitor the status bar for live capture activity during sessions.
- 5Review captured memory through the sidebar or CLI queries as needed.
Works with
Cortex: pros & cons
Pros
- +Zero-config background operation with automatic context injection
- +Multi-layer memory structure that balances recency and depth
- +Broad compatibility with major AI coding tools via MCP
- +Free tier options including local processing without API keys
Cons
- –Primary extraction quality improves only with paid or rate-limited API keys
- –Vector search features remain incomplete in the current version
- –Requires VS Code or compatible MCP environment for full functionality
Frequently asked questions
No, basic local pattern matching works without any key, though advanced extraction benefits from providers like Gemini.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar agents
Other coding & dev options worth comparing.
