> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tldt.me/llms.txt
> Use this file to discover all available pages before exploring further.
# How TLDT works
> Understand the boundary between your evidence, published Memory, and the public conversation.
TLDT uses a simple publication boundary: you provide evidence, you review the context, and visitors can use only the published version.
## The three layers
| Layer | What it means | Who can use it |
| --------- | --------------------------------------------------------- | ------------------------------------------------------------------ |
| Raw | Your source material, such as a CV, notes, or transcript. | You; selected public Raw excerpts can support a public answer. |
| Derived | A concise summary linked to Raw source paths. | You review it; published public Derived files can support answers. |
| Published | An immutable version chosen for the public runtime. | Public Chat, Fast Check, and pinned simulations. |
Editing a draft never silently edits a live profile. Publish a new version and activate it when you are ready.
## What the public assistant can do
TLDT can answer questions grounded in published Memory, explain a supported interpretation when you enable it, and use bounded public research for eligible validated conversations. Answers may show provenance labels such as **grounded**, **interpretation**, or **world knowledge**.
TLDT does not invent personal history, expose Private Memory, reveal source paths or internal traces, or accept a job, salary, schedule, or other commitment for you.
## Why this matters
The profile is designed for recruiter conversations, not as a generic chatbot or an automated hiring decision. It gives visitors useful context while keeping personal facts, boundaries, and publication decisions under your control.
When a fact matters, add it to Memory with the date, scope, actor, and source that support it. This gives TLDT something precise to use and gives you something precise to review.