> ## Documentation Index
> Fetch the complete documentation index at: https://docs.orka.sh/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> Connect your AI agents to your company's documents with a managed RAG pipeline.

Orka is a RAG platform that connects your AI agents to your company's documents. You upload files, configure how they're retrieved and reasoned over, and query them through an API or chat interface.

## Core Concepts

**Workspaces** — Isolated environments. All resources (agents, datastores, API keys) belong to a workspace. Each workspace has its own billing and team members.

**Datastores** — Collections of documents. You upload files and Orka chunks, embeds, and indexes them automatically. Agents search datastores at query time.

**Agents** — AI assistants with a system prompt and one or more connected datastores. Each agent has its own retrieval, reranking, and generation configuration.

**Domain Logic** — Business rules, metric definitions, and domain-specific terms that shape how agents interpret and answer queries.

## Quick Example

Send a query to an agent:

```bash theme={null}
curl -X POST https://orka.sh/api/v1/agents/{agent_id}/chat \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "message": "Which contracts are high risk?"
  }'
```

```json theme={null}
{
  "conversation_id": "3537b4ce-9445-4a24-b7a4-3d58bd1d5937",
  "response": "Per the High-risk contract definition, the following contracts qualify as high-risk...",
  "agent_id": "a33417bf-8577-42fe-a648-5b5fa18045d4",
  "processing_time_ms": 35330,
  "citations": [
    {
      "id": 1,
      "documentName": "Contract_TechFlow_MSA_2025.pdf",
      "pageNumber": 4,
      "excerpt": "CloudNest's total aggregate liability under this Agreement shall not exceed TWO MILLION DOLLARS ($2,000,000)...",
      "score": 0.42
    }
  ],
  "applied_definitions": [
    {
      "definition_name": "High-risk contract",
      "definition_type": "rule",
      "source": "workspace"
    }
  ]
}
```

Pass the returned `conversation_id` in follow-up requests to continue the conversation.

## Next Steps

* [Quickstart](/getting-started/quickstart) — Set up your account, create an agent, and send your first query