User Guide
==========

Getting Started
---------------

After installing **JupyDeep** for the first time, you'll need to configure a few settings to get the engine running:

1. Locate Settings
~~~~~~~~~~~~~~~~~~

Open the JupyterLab Settings Editor and search for options prefixed with ``jupydeep``.

2. Configure Your LLM
~~~~~~~~~~~~~~~~~~~~~

The agent engine requires an LLM backend to function. Set up your model provider 
(e.g., OpenAI) by following the configurations below (**Required**).

.. image:: _static/images/jupydeep_llm.png
   :scale: 80%
   :align: center

|

.. note::

   While OpenAI is used here as an example, any other supported LLM provider you have access to will work perfectly.
   Please refer to the available models listed in the `Pydantic AI (KnownModelName Section) <https://pydantic.dev/docs/ai/api/models/base/>`_.
   You can also configure local proxies—such as ``Ollama`` running on your local machine or network.

3. Configure Your MCP
~~~~~~~~~~~~~~~~~~~~~

To maximize productivity within the Jupyter workspace, it is highly recommended 
to configure the Jupyter MCP using the guidelines below (**Recommended**).

.. image:: _static/images/jupydeep_mcp.png
   :scale: 80%
   :align: center

|

.. warning::

   Ensure the URL prefix and port match those in your browser's address bar.

4. Meet Your Built-in Agents 
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Once your LLM is connected, two default agents will instantly populate your 
`jupydeep.agent` panel.

.. image:: _static/images/jupydeep_agent.png
   :scale: 80%
   :align: center

|

- 🐵 **Wu Kong:** Built for heavy-lifting code execution and technical problem-solving.

- 🐴 **Monk Tang:** Built for high-level tactical planning and orchestration.


.. tip::

   You can easily customize their skillsets and capabilities to fit your needs. 
   If the agents do not appear, try refreshing the page.


5. Global Setting 
~~~~~~~~~~~~~~~~~

This section primary configures the global behavior of the engine and agents. 
This includes settings for the default agent and model, token usage limits, 
tool-call timeouts, and backend configurations.

.. image:: _static/images/jupydeep_global.png
   :scale: 75%
   :align: center

|

.. warning::

   Some of these parameters are currently placeholders and have not been implemented yet.