Workspaces in Unli.ai

What is a Workspace?

A workspace is a dedicated environment for organizing and building AI-powered knowledge bases. It provides:

• Document isolation
• Focused AI knowledge management
• Dedicated storage and processing
• Unique configuration for your specific project or use case

Key Characteristics of Workspaces

• Isolated Environment: Documents and AI configurations are specific to each workspace
• Standalone Knowledge Base: Each workspace maintains its own set of documents and AI models
• Independent Configuration: Separate vector stores, embedding models, and settings

Creating a New Workspace

Step-by-Step Guide

1. Access Workspaces
   • Open the sidebar menu
   • Click on "Workspaces" section
2. Create Workspace
   • Click the "New Workspace" button
3. Configure Workspace
   • Name: Provide a clear, descriptive name
   • Description: (Optional) Add context about the workspace's purpose
4. Vector Store Configuration
   • Choose your vector store:
     • Managed Vector Store (Free plans)
     • Pinecone Vector Store (Paid plans)
   • Note: This selection cannot be changed after initial setup
5. Create Workspace
   • Click "Create Workspace" button

Post-Creation Configuration

Pinecone Vector Store

• If you chose Pinecone, you'll be prompted to enter your Pinecone credentials
• Action Required: Have your Pinecone API key ready
• Navigate to Pinecone dashboard to obtain your API key if needed

Embedding Model Configuration

• Select an embedding model:
  • OpenAI
  • Anthropic
  • Cohere
• Note: API keys are required for each embedding model

Language Model (LLM) Configuration

• Select a Language Model:
  • OpenAI
  • Google
  • Anthropic
• Requirement: Bring your own API keys

Workspace Use Cases

Example Scenarios

• Research Project: Dedicated workspace for academic literature
• Customer Support: Workspace for support documentation
• Product Development: Workspace for technical specifications
• Legal Compliance: Workspace for regulatory documents

Best Practices

1. Focused Scope: Create separate workspaces for distinct projects or knowledge domains
2. Consistent Naming: Use clear, descriptive names
3. Regular Maintenance: Periodically review and organize documents
4. Security Considerations: Use workspaces to maintain document confidentiality

Workspace Limitations

• One vector store per workspace
• Vector store type cannot be changed after initial setup
• Document isolation prevents cross-workspace searching

Managing Workspaces

Common Actions

• Create new workspaces
• Switch between existing workspaces
• Delete workspaces
• Manage workspace members (in team plans)

Performance and Scaling

• Each workspace has independent computational resources
• Scaling and performance depend on your selected plan
• Paid plans offer enhanced workspace capabilities

Security and Access

• Workspaces provide natural document segmentation
• Team and enterprise plans support granular access controls
• Ensures sensitive information remains compartmentalized

Getting Started Checklist

• Determine workspace purpose
• Choose an appropriate name
• Select vector store
• Obtain necessary API keys
• Select embedding model
• Select language model
• Upload initial documents
• Configure AI settings