Wayne Lundall

Ah, I see! Yes, I've seen discussions about using databases for AI context/memory management. While databases can be powerful for certain use cases (like chatbots needing to maintain conversation history across sessions, or multi-user systems requiring concurrent access to shared context), our current setup with markdown files in git is actually ideal for our use case because:

Documentation as Code

Our context is primarily development documentation

Lives naturally alongside the code

Benefits from git's version control

Easy to review and modify

Simplicity

No extra infrastructure needed

Files are human-readable

Direct editing in IDE

Clear structure in dev/

Collaboration

Changes tracked in git

Easy to review and merge

History preserved

Works with existing dev workflow

The file-based approach aligns well with software development best practices of keeping documentation close to code and treating it as a first-class citizen in version control.

Thanks for bringing up the question - it's a good validation that our current approach is appropriate for our needs!