KriyaOS Documentation

Build agents that answer, act, and integrate.

KriyaOS helps you create agents, connect knowledge and Kriyas, test their behavior, publish them, and integrate them into customer experiences or your own systems.

Start HereCreate your first agentUnderstand the two creation paths and choose the one that fits you best.CapabilitiesChoose what your agent can doMatch bookings, catalog recommendations, custom tools, and knowledge to the right capability.DevelopersIntegrate the APIChoose Public or Secure Chat, prepare the agent, and make the first request.

What You Can Do In KriyaOS

KriyaOS is a place to create, test, and publish AI agents for real work.

If you are new here, the simplest way to think about KriyaOS is:

KriyaOS helps you turn an idea like “I want an assistant that answers questions, helps customers, or represents my business” into something you can actually configure and use.

With KriyaOS, you can:

• create an agent for a business or workflow
• describe how that agent should behave
• add knowledge the agent can use when answering
• enable Kriyas for bookings, catalog recommendations, and custom tools
• test the draft before it goes live
• publish a version that other people can use
• integrate a published agent using Public Chat or Secure Chat APIs
• support follow-up conversations through channels like WhatsApp when appropriate

Choose Your Path

Pick the path that matches what you want to do first.

Path 1: Getting Started

Best if you are new to KriyaOS and want the core flow.

What Is KriyaOS?
What Is an Agent?
How to Create an Agent
Testing Your Agent

Path 2: Add Capabilities

Best if the agent needs to take actions, not only answer questions.

Choose Capabilities
Appointment Booking
Custom MCP

Path 3: Deploy & Integrate

Best if the agent is ready for customers or another application.

Choose an Integration
API Integration Quickstart

Explore The Docs

Getting StartedLearn what KriyaOS is and create your first agent with the right setup path.AgentsUnderstand what an agent is, how drafts work, and when a version is ready to publish.KnowledgeSee how agents become more useful with real information and what to expect after upload.APIsUse public or secure chat APIs to integrate published agents into your product and workflows.Capabilities (Kriyas)Give agents controlled actions such as booking appointments, recommending catalog items, and calling MCP tools.PlaygroundTest your agent, review replies, and refine the setup before or after publishing.IntegrationsFind the right place to learn how published agents support customer communication across channels, including WhatsApp when it fits your business.WorkspaceManage members, invitations, API keys, and billing for your KriyaOS workspace.CredentialsStore provider API keys securely and understand primary, fallback, and BYOK model behavior.TroubleshootingGet help when agent behavior, knowledge readiness, or setup expectations do not match.

Kriya Agents in Action

Real-world scenarios of how businesses use KriyaOS to automate interactions and improve customer experience.

Barbers & SalonsSee how busy shops can answer common booking and service questions more smoothly.Clinics & HealthcareExplore how clinics can handle routine patient questions with more clarity and consistency.Restaurants & CafesLearn how restaurants and cafes can guide guests before they visit, order, or book.Coaches & ConsultantsSee how service businesses can explain offers and guide potential clients more clearly.Local Service ProvidersUnderstand how local businesses can reduce repetitive inquiries and respond faster.

What An Agent Means Here

In KriyaOS, an agent is your AI assistant.

It is not just a chat box. It includes:

• a name and purpose
• instructions for how it should respond
• the model it uses
• any knowledge you connect to it
• a draft version and, when you are ready, a live version

If you want the full explanation, read What Is an Agent?.

Suggested Path

If you are new to KriyaOS, this is a strong place to begin:

1. How to Create an Agent
2. Draft vs Published
3. Choose Capabilities
4. Testing Your Agent
5. API Integration Quickstart

What These Docs Focus On Right Now

The documentation follows the full path from setup to an integrated, published agent:

create → add knowledge and capabilities → test → publish → integrate