A vector store indexes and stores vector embeddings (the vectorial representation of documents) for fast retrieval and semantic search. Embeddings are generated by AI models, in the context of machine learning these features represent different dimensions of the data that are essential for understanding patterns, relationships, and underlying structures.

You handle the vector stores in Administration > Artificial Intelligence > Embeddings > Vector Stores

At the time of writing, LogicalDOC just supports MariaDB as vector store; in the future, more vector store providers will be made available.

It is therefore required to insert the connection details to a MariaDB (version 11.8 or greater) database; you may use the wizard icon to get some help in composing the connection Url.

A robot is an intelligent agent designed to understand user questions and provide meaningful answers.

Robots act as an interface between the user and the Natural Language Processing (NLP) engine, using trained models to classify queries and extract key information. The platform provides a default robot is named Coach, but you may create your ones dedicated to specific areas.

How Robots Work

Each robot is configured with two core NLP models:

• Classifier: categorizes the user’s question into a specific action or intent (e.g., GETDOC, SEARCHFILE, UNKNOWN).
• Tokens Detector: extracts specific values from the text, such as document IDs or filenames.

When a user asks a question, the robot:

• Classifies the sentence using the classifier.
• Extracts tokens using the tokens detector.
• Executes a matching automation script (called an “answer”) associated with the identified category.
• If the classification or token extraction fails, a default fallback answer is used.

Robot Configuration

Robots are configured through the Robot Management Interface.

The most relevant aspect of a robot's configuration is the Answer section, where each category (e.g., GETDOC, SEARCHDOC, SEARCHFILE) is mapped to an Automation script. These scripts define how the robot responds to a user query once the classifier and tokens detector have done their job.

Using these scripts, you can:

• Retrieve and open documents by ID
• Perform keyword-based full-text searches
• Look up files by name
• Handle unknown queries gracefully

These answers are stored as Automation scripts, allowing advanced conditional logic, data access, and dynamic rendering of results. 

Training AI models is a computationally intensive and time-consuming process, often requiring specialized hardware and large datasets. For this reason, not all models can be trained directly within LogicalDOC.

To simplify adoption, the system allows users to install and use pre-trained models, which are ready to operate without additional training.
These models can be downloaded and imported into the platform, enabling advanced features such as document tagging and language detection with minimal setup.

The import process is straightforward:

1

Enter the models section

Go to Administration > Artificial Intelligence > Models

2

Download the Model file

The model to import must be a file exported from another installation or downloaded from the LogicalDOC's download center. Here are the currently available pre-trained models:

3

Import the Model file

Once you have downloaded your model file, Click on the Import button.

Once the Upload window appears click on Upload and select the downloaded model.

The model will appear in the list of the available models.

Embeddings are vectors representing entire documents or fragment of them into a continuous vector space. This numerical representation of is required to efficiently infer similitudes between documents and implement features like Semantic Search.

This means that LogicalDOC must calculate all these embeddings for the documents in your repository and save them into the Vectors Store, whose setup is a requirement.

Embedding Schemes

The process of calculating an embedding of a document is not unique, but depends on what embedding model you use.

In Administration > Artificial Intelligence > Embeddings, you can handle different embedding schemes, each one telling LogicalDOC how to process the documents with a specific embedding model.

When you create a new scheme by clicking on Add embedding scheme, you will be required to specify one of the available embedding models.

At the time of writing, you can choose among the Embedder models directly coded in LogicalDOC itself or one of the embedding models available in ChatGPT.

Settings common to all embedding models are:

• Batch: The maximum number of Documents written to the vector store in a single operation.
• Chunks batch: How many chunks gets added into the vector store at the same time.

Settings specific to ChatGPT model are:

• Model Spec.: name of the embedding model to use, e.g.: text-embedding-3-small
• Vector size: must match the exact size of the embeddings produced by the chosen model, in case of text-embedding-3-small the size is 1536
• API Key: your API key provided by ChatGPT

For more information about ChatGPT embeddings, please refer to https://platform.openai.com/docs/guides/embeddings

Settings

Click on the Settings button to see configuration parameters that regulate how the task works.

• Include patterns: which document type should be embedded. If left empty, all documents will be included by default.
• Exclude patterns: which document type should not be embedded. For example to exclude all documents with .png extension, you can insert *.png in the field. 
• Batch: number of Documents to be processed together by the Embedder task. 
• Sorting: Determines the order in which pending documents are embedded (e.g., prioritizing newer files or smaller files to optimize throughput).
• Custom Sorting: Possibility to define a custom sorting logic.
• Threads: how many parallel "workers" are running at the same time to process the embedding queue.

As vectors are calculated and saved into the vector store, you can see this in the counter and in the Embeddings panel.