memory

To keep coding context persistent across sessions, you can use automated memory storage to save project decisions, code changes, and notes. This allows your coding assistant to recall and summarize the exact context when you start a new session.

Automatic summarization works by synthesizing raw loaded memories into a human-readable summary immediately after each retrieval. This provides quick understanding of past project context without needing to manually parse individual memory entries.

Yes, you can leverage Graphiti MCP with a cache fallback to ensure memories remain available even if the endpoint is temporarily unavailable. This resilient storage approach prevents context loss during endpoint connectivity issues.

To group and tag saved memories for feature branches, you use the add command with optional groupings and tags. The group ID is automatically derived from the project folder path, while tags help categorize specific file changes or design decisions.

Yes, you need to configure a Graphiti endpoint to enable persistent memory recall. The endpoint is read from the environment variables inside your project's configuration directory, which is automatically written during initialization.

The best way to recall past code changes is to load memories using an optional query and limit parameters. The system retrieves the stored context and automatically produces a summarized overview of the relevant project history.