Workspace Configuration

Workspaces in Inquira Labs serve as dedicated environments for organizing and managing documents, interactions, and configurations. They allow you to maintain structure and access control across different projects, teams, or use cases.

1. What Are Workspaces?

Workspaces are isolated containers where:

• Documents are stored and managed.
• Chats and interactions are conducted.
• Specific configurations are applied, ensuring tailored behavior for the AI.

Each workspace can have its own settings, permissions, and data, making it a flexible solution for handling multiple projects or teams.

2. Key Configuration Elements

Workspace Creation

• When creating a workspace, you can define a unique name and description for identification.
• A workspace ID is automatically generated, serving as a unique identifier for API interactions.

Default Settings

• Each workspace inherits default global settings for:
  • LLM model (Language model used for responses)
  • Embedder model (Used for text vectorization)
  • Vector database (Storage and retrieval of vectorized documents)
• These defaults can be overridden at the workspace level for tailored configurations.

3. Customizable Options

1. Workspace LLM Configuration

• Select the language model (LLM) for chat responses.
  • Hosted options like OpenAI models.
  • Local LLMs hosted on your infrastructure.
• Adjust parameters such as model temperature for more dynamic or deterministic outputs.

2. Document Management

• Upload and organize documents in the workspace.
• Supported formats include PDFs, Word documents, and plain text files.
• Assign metadata to documents for better categorization and searchability.

3. User Roles and Permissions

• Assign users to workspaces with specific roles:
  • Admin: Full access to manage documents, chats, and settings.
  • Manager: Moderate access to manage users and review activity but limited system control.
  • Default User: Can view and interact with documents but cannot modify settings.

4. Chat and Query Modes

• Configure how the chatbot interacts with users:
  • Chat Mode: The chatbot answers all queries, regardless of their relevance to workspace documents.
  • Query Mode: The chatbot responds only to questions directly related to workspace documents, ensuring context-sensitive interactions.

5. Access Restrictions

• Limit access to specific IP addresses or domains to secure the workspace.
• Configure allowed referrers for embedded chat widgets.

4. Advanced Features

Dynamic LLM and Embedder Use

• Workspaces can override the global LLM and embedder configurations to use different models suited for specific needs.
• For example, one workspace can use OpenAI’s GPT-4, while another uses a locally hosted model.

Use Cases

1. Project Management
   Create a separate workspace for each project, storing relevant documents and configuring the chatbot to answer only project-specific queries.
2. Team Collaboration
   Assign different roles to team members, ensuring secure and structured collaboration.
3. Customer Support
   Configure workspaces for specific clients or products, allowing tailored chatbot responses and document access.