Files

Files are the primary way information enters Karyam.

They serve as the foundation for:

Knowledge ingestion
Document processing
Retrieval-Augmented Generation (RAG)
AI context enrichment
AI flow automation

Files can be uploaded manually or automatically through Listeners.

Why Files Exist

AI systems become significantly more useful when they understand your organization’s data and documents.

Examples include:

Policies
Standard operating procedures
Contracts
Technical documentation
Knowledge articles
Reports
Emails
Images
Audio transcripts

Files provide the entry point for this information.

Supported Use Cases

Files are commonly used for:

Enterprise Knowledge

Upload documents to make them searchable by AI systems.

Examples:

HR Policies
VPN Access Guidelines
Employee Handbooks
Compliance Documents

Document Processing

Files can trigger AI flows for extraction, classification, or analysis.

Examples:

Invoice Processing
Contract Review
Resume Screening
Report Summarization

AI flow Automation

Files can act as triggers for automated business processes.

Example:

Invoice Uploaded
      ↓
Folder Listener
      ↓
AI Flow
      ↓
Approval AI flow
      ↓
ERP Update

File Upload Methods

Files can enter Karyam through multiple mechanisms.

Manual Upload

Users can upload files directly through the platform interface.

Common examples include:

PDFs
Markdown files
Text files
Images
Audio files

Folder Listeners

Folder listeners automatically process files that appear in monitored directories.

Example:

/shared/contracts
/shared/invoices
/shared/policies

Email Listeners

Attachments received via email can automatically become part of AI flows.

Example:

contracts@company.com
invoices@company.com

HTTP Listeners

External applications can upload files using APIs or webhooks.

This enables integrations with:

Internal systems
CRMs
ERPs
Document management platforms

Files and Knowledge

Files become enterprise knowledge when combined with:

Embedding Models
Embedding Listeners
Vector Databases

The ingestion pipeline looks like:

File Upload
     ↓
Embedding Listener
     ↓
Artifact Generation
     ↓
Chunking
     ↓
Embeddings
     ↓
Vector Database

Once processed, the content becomes available for semantic retrieval.

Files and RAG Runs

Knowledge ingestion operations are visible through:

AI INFRA
└── RAG Runs

RAG Runs provide visibility into:

Artifact generation
Chunk creation
Embedding generation
Vector storage
Retrieval activity
Processing failures

Files and AI Systems

Files do not directly participate in reasoning.

Instead, they provide the context required for intelligent decision making.

Example:

Employee asks:
"My VPN access expired."

Agent
   ↓
Retrieves VPN Policy
   ↓
Determines Approval Requirement
   ↓
Continues AI flow

Without files, the AI system would lack organizational context.

Relationship to Other Concepts

Files
   ↓
Listeners
   ↓
Embeddings
   ↓
Vector Databases
   ↓
Retrieval
   ↓
Agents

Files form the starting point of Karyam’s knowledge infrastructure.

The Karyam Philosophy

AI systems should operate using the same knowledge your organization uses every day.

Files bridge the gap between business information and AI reasoning.