Unli.ai docs

Appearance

Getting Started with Unli.ai ​

Transform Your Data into AI Knowledge ​

Welcome to Unli.ai! This guide will help you quickly set up and start using our enterprise-grade RAG (Retrieval-Augmented Generation) infrastructure.

Unli.ai allows you to transform any document into AI-ready knowledge while maintaining complete control over your data. Our platform deploys in minutes and provides everything you need to create sophisticated AI knowledge bases from your existing content.

Prerequisites ​

Before getting started with Unli.ai, you'll need:

  1. Vector Database:

    • A Pinecone account (currently the only supported vector database)
    • Coming soon: Support for Chroma, Weaviate, and others

  2. Storage for Documents:

    • An S3-compatible storage solution for uploading files or scraping websites:
      • AWS S3
      • Cloudflare R2 (recommended)
      • Any S3-compatible storage provider

  3. Embedding Models:

    • Access to one of the following embedding model providers:
      • OpenAI models (recommended)
      • Anthropic models
      • Cohere models

Creating Your First Project ​

A project in Unli.ai is a dedicated workspace where you can isolate and organize documents to create AI-based knowledge. Content added to a project can be aggregated and searched together.

Step 1: Sign Up for Unli.ai ​

  1. Navigate to Unli.ai's website
  2. Click the "Sign Up" button
  3. Complete the registration process

Step 2: Create a New Project ​

  1. From your dashboard, click "Create New Project"
  2. Give your project a name and description
  3. Configure project-specific settings if needed
  4. Click "Create"
  5. Configure your:
    • Vector Database connection (Pinecone)
    • Storage provider (S3, R2, etc.)
    • Embedding model provider (OpenAI, Anthropic, or Cohere)

Step 3: Add Content to Your Project ​

You can add content to your project in several ways:

Upload Documents ​

  1. Navigate to your project
  2. Click "Add Content" → "Upload Documents"
  3. Select files from your computer
  4. Click "Upload"

Scrape Websites ​

  1. Navigate to your project
  2. Click "Add Content" → "Scrape Website"
  3. Enter the URL(s) you want to scrape
  4. Configure scraping options
  5. Click "Start Scraping"

Connect to Data Sources ​

  1. Navigate to your project
  2. Click "Add Content" → "Connect Data Source"
  3. Select your data source type
  4. Configure connection settings
  5. Click "Connect"

Step 4: Process Your Content ​

  1. Once content is added, Unli.ai will automatically begin processing:

    • Converting documents to text
    • Chunking content appropriately
    • Creating embeddings using your selected model
    • Storing vectors in your Pinecone database

  2. You can monitor progress in the "Processing" tab

Step 5: Test Your Knowledge Base ​

  1. Navigate to the "Test" section of your project
  2. Ask questions related to your content
  3. Review the responses and cited sources
  4. Adjust settings as needed to optimize performance

Next Steps ​

  • Fine-tune your chunking strategy: Adjust chunk size and overlap for better results
  • Customize your retrieval settings: Modify similarity thresholds and result count
  • Set up API access: Generate API keys to integrate your knowledge base with other applications
  • Monitor usage: Track usage metrics and optimize your configuration