brew tap asakin/llm-primer && brew install llm-primer
primer start
primer # attach to a warm session
Built this for my own workflow — I'm running LLM wikis
(llm-context-base) and switching context between different agent roles
constantly. The startup latency was killing the flow. Happy to answer
questions.
I've been running a personal wiki system for a while. When Karpathy posted his LLM Wiki pattern I realized I'd been building toward the same thing, so I adapted what I had into a git template.
It adds a metadata standard (every file has a header so the AI can scan by summary without loading full documents), a lint pass for stale content, and some other scaffolding on top. Still just a folder of markdown files - no app, no account, you own everything.
Happy to answer questions about any of the design decisions.
Main additions on top of the pattern: a training period where the AI learns how you work over 30 days then gets quieter over time, a metadata standard so files are queryable by summary, and a lint pass for stale content and context loading optimization. Never have to design a taxonomy upfront.