Skip to main content
A notebook is a named collection of learning records. Each record is an output saved from another module — a solved problem, a generated question set, a research report, or a co-writer document. Notebooks give you a persistent, browsable history of your work that other modules (guided learning, idea generation) can read back as source material.

Cross-module collection

Accepts records from solver, question generator, deep research, and co-writer.

Custom appearance

Each notebook has a configurable color and icon for quick visual identification.

Used by guided learning

Guided learning reads notebook records to generate personalized learning paths.

Used by idea generation

Automated IdeaGen extracts knowledge points directly from notebook records.

​
How to use

1

Open the notebook page

Navigate to http://localhost:3782/notebook.
2

Create a notebook

Click New Notebook. Set a name, optional description, color, and icon.
3

Save records from other modules

After completing any activity — solving a problem, generating questions, finishing a research session, or editing in co-writer — click Add to Notebook and select the target notebook.
4

Browse and manage records

Open the notebook to view all saved records. Records are sorted and tagged by type and knowledge base.
5

Use notebooks in other modules

In guided learning and idea generation, select a notebook (or multiple notebooks) as the source. The module reads the records automatically.

​
Creating and managing notebooks

Each notebook has the following properties:
PropertyDescription
NameRequired. Identifies the notebook in the sidebar and module pickers.
DescriptionOptional freeform notes about the notebook’s purpose.
ColorHex color used for the notebook’s visual chip.
IconIcon displayed next to the notebook name.
You can create, edit, rename, and delete notebooks from the notebook page. Deleting a notebook removes the metadata and record list; it does not delete the original output files from other modules.

​
Record types

Records are automatically categorized by their source module.
Record typeSource moduleWhat is stored
SolveSmart solverFinal answer, solution chain, citations
QuestionQuestion generatorGenerated question set and validation report
ResearchDeep researchResearch report path and metadata
Co-writerCo-writerDocument content and edit history reference
When you click Add to Notebook inside any module, DeepTutor records a reference to that session’s output along with a timestamp, source type, and the knowledge base used (if applicable).

​
Cross-module integration

​
Guided learning

The guided learning module reads selected notebook records to plan a personalized learning path. You can select records from one or multiple notebooks in the session setup screen. The LocateAgent analyzes the combined records to identify 3–5 progressive knowledge points for the session.

​
Idea generation

Automated IdeaGen’s MaterialOrganizerAgent processes notebook records as its primary input. The more varied the record types in the notebook, the broader the pool of knowledge points the agent can discover. You can select a single notebook per idea generation session.
Maintain separate notebooks by topic or project to keep idea generation and guided learning sessions focused. A notebook with mixed unrelated records produces less coherent output.

​
Categories and tags

Records are automatically tagged with:
  • Type tag — solve, question, research, or co-writer.
  • Knowledge base tag — the name of the knowledge base used in the session (when applicable).
These tags are used for filtering within the notebook view and are passed to downstream modules so agents can understand the provenance of each record.

​
Visual customization

Each notebook can be assigned a custom color and icon from the notebook creation or edit dialog. These settings are purely cosmetic — they make it easier to identify notebooks at a glance in the module pickers throughout the application.

​
Data storage

Notebook metadata is stored in data/user/notebook/. The notebook files contain record lists and metadata; the actual content (reports, solutions, etc.) remains in the module-specific output directories. 
data/user/
├── notebook/              # Notebook metadata and record lists
├── solve/                 # Solver outputs referenced by notebook records
├── question/              # Question outputs referenced by notebook records
├── research/              # Research reports referenced by notebook records
└── co-writer/             # Co-writer documents referenced by notebook records
Notebooks and their records persist across sessions. You can close and reopen DeepTutor and all notebooks remain accessible. If you move or delete the underlying output files in data/user/, the notebook record will still exist but the linked content will no longer be accessible.

Build docs developers (and LLMs) love

Get started for freeTalk to us