Skip to content

Knowledge Hub

Introduction

The Altimate Knowledge Hub empowers your organization to centralize and leverage its tribal knowledge and best practices, directly within your AI workflows, particularly in your IDEs like Cursor, Github Copilot, Windsurf. By providing this curated information as verified Context, you can achieve more reliable and relevant AI-generated outputs, curb hallucinations, and move beyond generic code generation to solutions tailored to your specific organizational needs.

LLMs rely on outdated or generic information about the libraries or components you use. The "best" LLM doesn't know about any internal knowledge or best practices that have been built through years of collective human effort.

This guide will walk you through accessing, creating, and utilizing your knowledge within the Altimate Knowledge Hub.

What You Can Do

  • Create Unified Knowledge: Keep all your team's best practices, code standards, and guides in one spot
  • Leverage Expert-Built Templates: Start with pre-built templates created by Altimate, drawing on our years of in-house data technology expertise and best practices refined through working with multiple global technology giants.
  • Contextualise your AI Workflows: Each knowledge doc gets a unique link which can be used in your IDE of choice to get AI to understand your team's specific context.

When you open the Knowledge Hub, you'll see:

knowledge-hub-img

  • My Knowledge: This is where your team's custom knowledge documents live. If it's your first time, this might be empty.

  • Altimate Knowledge Templates: A list of ready-to-use documents (e.g., "DataOps Best Practices," "PySpark Best Practices"). You can:

    • View Details: See what's inside a template.
    • Fork: Copy a template to "My Knowledge" to customize it.
  • Create New Knowledge button: Lets you start a new knowledge document from zero.

Creating New Knowledge

1. Forking (Copying) a Template:

This is the fastest way to start.

  1. Find a template you like in "Altimate Knowledge Templates."
  2. Click Fork.
  3. Give your new document a clear name (e.g., "Our Team's Python Standards").

  4. Click Fork again in the pop-up.

    Your new document will now appear in "My Knowledge."

2. Creating from Scratch:

  1. Click the Create New Knowledge button (usually top-right).
  2. Follow the steps to set up your knowledge.

You can paste upto 1MB of text

Using Available Knowledge

This is where Knowledge Hub shines. You tell your IDE agents like Cursor, Github Copilot, Windsurf, Cline, etc. to use your team's specific knowledge for context. First, get the link

Get the Link:

  • Go to "My Knowledge" in the Altimate Knowledge Hub. (https://.app.getaltimate.com/knowledge-catalog)

  • Open the document you want the AI to use by clicking "View Details."

  • Copy its Link (e.g., https://apimi.getaltimate.com/knowledge_...).

Private Links - Request Only. Access your Internal organisation knowledge as locally a hosted knowledge hub. You can contact us over Slack or Chat.

For Cursor Users

Use it in Cursor:

  • In Cursor, go to Agent Mode (Cmd + I) or (Ctrl + Shift + I)
  • Type @doc select Docs, scroll to the bottom of the list and click on add a new doc
  • Paste the copied link.
  • Provide a name for your doc (my-kb)
  • Now you can reference your knowledge as @my-kb in your tasks.

Example - 

Refactor this @model.sql file using my best practices @my-kb
Follow this tutorial to use your knowledge in Cursor -

For Github Copilot Users

Use it in Github Copilot:

  • In Github Copilot, go to Agent Mode (Cmd + Control + I) or (Ctrl + Shift + I)
  • Type #fetch and paste the copied link
  • Approve Github Copilot to access the link
  • (Optional) If you don't want to explicitly approve every time, you can add this link to your Trusted Domains in VSCode

Example -

Refactor this @model.sql file using my best practices #fetch <link>

For Cline Users

Use it in Cline:

  • In Cline, type @

Example - 

Refactor this @model.sql file using my best practices @<link>

For Windsurf Users

Use it in Windsurf:

  • In Windsurf, type @web

Example -

Refactor this @model.sql file using my best practices @web <link>

Knowledge Engine

Enterprise Feature

Knowledge Engine is available in the Enterprise edition only. Contact us over Slack or Chat to upgrade.

What is Knowledge Engine?

Knowledge Engine is an intelligent document processing system that transforms your uploaded documents into queryable knowledge bases. When you upload documents to a knowledge base, the Knowledge Engine:

  1. Parses the Document: Extracts text and processes embedded images
  2. Creates Retrievable Context: Indexes content for intelligent retrieval through AI agents
  3. Improves performance and efficiency: Instead of sending full documents as context to AI agents - which consumes a lot more tokens and bloats overall context quality, Knowledge Engine surfaces only the relevant knowledge snippets as context

Why Knowledge Engine

Knowledge engine goes beyond simple text storage by:

Create Knowledge Base

  • Processing Multiple Formats: Upload PDF, Word (.docx), Markdown (.md), or text files
  • Understanding Multimodal Content: Extracts insights from both text and embedded images (diagrams, charts, screenshots)
  • Document Editing: Update and edit your uploaded documents to keep knowledge current
  • Intelligent Context for Agents: Automatically provides relevant context to AI agents in your IDE through MCP integration

Knowledge Engine Components

The Knowledge Engine interface provides three key views:

1. Document Tab

Shows your uploaded document once it is parsed.

2. Learnings Tab

Shows the structured knowledge extracted from your document:

  • Learning name: Name given to a certain group of information within the document
  • Tags: Auto-generated categorization
  • Document: Name of the actual document
  • Retrieval Count: Track how often each learning is accessed

3. History Tab

Monitor document processing and versions:

  • Upload history with timestamps
  • Processing status (Uploading → Processing → Completed/Failed)
  • Version control for document updates
  • File metadata (filename, type, size, uploaded by etc)

Enabling Knowledge Engine for Your Datamate

To use Knowledge Engine with your Datamates in the IDE:

Enable Knowledge Engine for DataMate

  1. Navigate to DataMates in your Altimate platform
  2. Select or create a Datamate
  3. In the Datamate, locate the Knowledge Hub section
  4. Toggle On the Knowledge Engine option
  5. Select which knowledge bases your Datamate should have access to

Once enabled, your Datamate can automatically retrieve relevant context from your knowledge bases when you query it through the MCP server in your IDE.