Implicit Loading guarantees an AI client loads your 📝Augmentation System context before it can do anything else inside 📝Augmentation — MythOS enforces this at the MCP server level, rejecting every other tool call until the context has been pulled. It exists because a context step an AI can skip eventually gets skipped, and a relationship that depends on discipline isn't a relationship you can rely on. It's for anyone whose AI has ever started a session as a stranger.
Key Capabilities
- A mandatory first call — an MCP client must call for your augmentation context before any other MCP tool will run
- Per-key, time-boxed — the requirement is satisfied for 24 hours per connected API key, not re-required on every single tool call
- Full augmentation payload — the response includes your Mythos Soul, Style, Human, and Memory, formatted for the AI to read directly
- Individually addressable too — each memo is also available as its own MCP resource, for clients that want to fetch pieces separately
Getting Started
Nothing to configure — any MCP-connected client (Claude, Cursor, Claude Code, and others) hits this gate automatically the first time it tries to use your library each day.
FAQ
- Can an AI client skip this and go straight to reading my memos? No — every other MCP tool call is rejected with an explicit instruction to load context first, until the gate is satisfied.
- Do I have to trigger this myself? No — a well-behaved MCP client calls it automatically; the gate exists precisely so this doesn't depend on the human or the AI remembering.
- How often does the gate reset? Every 24 hours per connected API key — after that, the next tool call re-triggers the requirement.
- Does this replace the individual augmentation resources? No — both exist: the gate delivers the full payload at once, and each memo is also fetchable on its own as an MCP resource.
Related
- 📝Session Memory Protocol — the daily-memo layer that builds on this same context