Roadmap: which direction matters most to you?

[alexanderbering]

docs/ROADMAP.md lists the candidate directions. None are scheduled — development is self-funded and proceeds without external deadlines, so contributor and user interest genuinely moves things up the list.

The five clusters:

• Adapters & integration — more storage adapters (e.g. Redis caching layer), built-in embedding/LLM provider adapters, npx zenbrain init scaffolding
• Algorithms & scale — graph-aware retrieval combining Hebbian edges with semantic search, batch FSRS scheduling, episodic compression, event-driven consolidation
• Memory layers — a cross-context layer that the coordinator actually uses (today it is experimental and not wired in), and resumable short-term sessions
• Reproducibility & evaluation — a reproducible comparative benchmark suite against other open-source memory systems, with published methodology and raw results
• Production — encryption at rest, multi-tenant namespace isolation, federated memory

If one of these would unblock a real project of yours, say so below — a sentence about the use case is worth more than a +1. If you would rather build one: #56, #57 and #58 are open good-first-issue requests, and CONTRIBUTING.md has the workflow.