Kanbn Integrations Command
========================

The 'integrations' command manages markdown files that provide additional context for AI assistance using RAG (Retrieval-Augmented Generation).

Usage:
------
kanbn integrations [options]

Description:
-----------
Manages integration markdown files that enhance the AI context when using chat commands.
Integrations can be added from URLs or directly with content, and then used with the chat command.

The RAG-based approach:
- Stores integration content as markdown files in the .kanbn/integrations directory
- Uses vector embeddings to find the most relevant information for each query
- Automatically retrieves only the most relevant parts of integrations based on the user's question
- Combines knowledge from multiple integrations when needed
- Uses local Ollama models for embeddings when available
- Falls back to simple keyword matching when embeddings are not available
- Automatically converts HTML webpages to Markdown when adding from URLs

Options:
-------
--list                     List all available integrations
--add                      Add a new integration
--remove                   Remove an integration
--name <name>              Name of the integration (required for add/remove)
--url <url>                URL to download integration content from
--content <content>        Content for the integration
--help                     Show this help message

Examples:
--------
List all integrations:
  kanbn integrations --list

Add an integration from a URL:
  kanbn integrations --add --name github --url https://raw.githubusercontent.com/username/repo/main/README.md

Add an integration from an HTML webpage (automatically converted to Markdown):
  kanbn integrations --add --name docs --url https://js.langchain.com/docs/tutorials/rag/

Add an integration from content:
  kanbn integrations --add --name api-spec --content "# API Specification\n\nThis is the API spec..."

Remove an integration:
  kanbn integrations --remove --name github

Using integrations with chat:
---------------------------
To use integrations with the chat command, you can use the following options:

Use all available integrations:
  kanbn chat --with-integrations

Use specific integration(s):
  kanbn chat --integration github
  kanbn chat --integration github --integration api-spec

Example queries that leverage integrations:
  kanbn chat --integration game-systems --message "What stats are used in D&D 5E?"
  kanbn chat --with-integrations --message "How do I query users in our database?"
  kanbn chat --integration api-docs --message "What endpoints are available for user management?"

How RAG works with your queries:
1. When you ask a question, the system converts it into an embedding vector
2. It searches for the most relevant sections in your integrations
3. Only the most relevant content is included in the AI context
4. The AI uses this targeted information to provide more accurate answers
5. This approach allows for much larger knowledge bases without context limitations

These options make the AI more knowledgeable about your specific project needs
while maintaining efficiency by only including relevant information.