> ## Documentation Index
> Fetch the complete documentation index at: https://docs.mossx.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Project memory and knowledge map

> Add long-lived context to chat with project memory, the knowledge map, and the context ledger.

Desktop-only project intelligence helps the model understand your repo conventions and structure.

## Project memory

Save key conventions and notes, such as:

* Code style
* Release process
* Module boundaries
* Lessons learned

When you send, you can pick memory with `@@`. You can also turn on Memory Reference retrieval and inject it into the current turn.

<img src="https://mintcdn.com/mossx-docs/kz4nt5KFSKzkAZwe/images/desktop/project-memory.png?fit=max&auto=format&n=kz4nt5KFSKzkAZwe&q=85&s=203b0e812be1fd4fcef072a03c7ceede" alt="Project memory" width="100%" data-path="images/desktop/project-memory.png" />

Settings entry: **Settings → Project memory**.

## Project knowledge map

After AI scans the project, it builds an interactive structure map:

* File relationships
* API endpoints
* Module dependencies
* Incremental updates

<img src="https://mintcdn.com/mossx-docs/kz4nt5KFSKzkAZwe/images/desktop/project-map.png?fit=max&auto=format&n=kz4nt5KFSKzkAZwe&q=85&s=9446fff11fd09e41277b7631241ccd67" alt="Project knowledge map" width="100%" data-path="images/desktop/project-map.png" />

## Context ledger

It shows:

* Selected or inherited context sources
* Token / character estimates
* Freshness
* Attribution confidence

Use it to see which context this turn actually carries, so you do not blow the limit by accident.

## Note cards

Local notes can be attached in chat with `@#`. They fit checklists and short scratch notes.
