working-with-context

You can store project context for codebase analysis by onboarding your project to capture file inventory, tech stack, and module patterns, then reusing the stored context in subsequent analyses to reference existing architectural decisions.

Context-aware codebase analysis is the process of referencing stored project context, such as file structures and module patterns, to provide accurate recommendations. You need it when onboarding new projects or maintaining consistency across multiple analysis sessions.

To onboard a project and capture its tech stack, run the onboarding command with your project path and a project ID. This process inventories files, structure, and patterns, storing them locally and in Serena MCP for future contextual usage.

Yes, you can persist codebase context locally and in Serena MCP for collaborative projects. This allows team members to maintain and reference the same project context across different analysis sessions.

The best way to analyze new features using stored project context is to run the analyze command with your feature file and project ID. This leverages previously stored structural data to deliver recommendations that align with existing code patterns.

While Serena MCP is used for persisting data in collaborative projects, the context can also be stored locally. You can onboard and analyze codebase structures without Serena MCP, but collaborative context sharing requires it.