STARTUP.md
Runs once after cloning. Personalizes the template, fills in identity files, and confirms the system is ready.
Open guide
The lightweight AI operating system for building a personalized home for your AI: workflows, wikis, memory, templates, and repeatable systems.
CenterOS is a cloneable AI operating system. It turns a plain repo into a structured workspace where an AI can learn your preferences, follow your rules, run documented workflows, and maintain useful knowledge over time.
It is intentionally lightweight. The system uses Markdown files, simple folders, relative paths, and append-only logs so anyone can inspect it, change it, and move it between tools.
After cloning the template, open the repo in your AI coding harness and say:
Read STARTUP.md and follow its instructions.
The startup process asks a short questionnaire, replaces identity placeholders, and converts the public template into your own personalized AI operating system.
Runs once after cloning. Personalizes the template, fills in identity files, and confirms the system is ready.
Open guideLoads the AI's operating context for normal sessions after the first-run setup is complete.
View referenceExplains how workflows, wikis, memory, templates, and logs fit together as one operating system.
Read architectureRepeatable procedures live in workflows/. Each workflow documents what it needs, what it produces, and the steps the AI should follow.
Knowledge bases live in wikis/. Raw source material stays immutable while synthesized pages build a useful reference layer.
Durable memory lives in memory/. It stores durable preferences, decisions, feedback, and context inside the repo.
Reusable scaffolds live in templates/. They keep new workflows and wikis consistent.
Append-only logs record meaningful workflow runs and system changes so future sessions can see what happened.
SYSTEM_INDEX.md is the internal map of the personalized operating system after startup.
CONTEXT.md.LOG.md.CenterOS pushes deterministic work into scripts and repeatable procedures whenever practical. AI judgment is reserved for summarization, writing, nuanced classification, ambiguous interpretation, and decisions that actually need context.