Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kataven.ai/llms.txt

Use this file to discover all available pages before exploring further.

Kataven is a multi-tenant voice AI platform. You build agents, configure their tools and knowledge, hand them a phone number or embed them in a web page, and customers can talk to them — over PSTN or in a browser widget. The Hub API is the management surface. It’s what the dashboard at hub.kataven.ai talks to, and it’s what your own scripts, CI pipelines, or AI agents (Claude, etc.) use when they need to drive Kataven without clicking through the UI.

What you can do with the API

Manage agents

Create, configure, and version voice agents — system prompt, voice, LLM, tools, knowledge, the whole config.

Wire up tools

Browse the built-in tools, write custom ones, install third-party integrations.

Telephony

Connect Twilio / Plivo / Vobiz, assign phone numbers, originate outbound calls.

Outbound campaigns

Bulk-dial from CSV with cap-aware dispatch, pause/resume, real-time metrics over SSE.

Embeddable widget

Drop a launcher into any web page; manage client keys, branding, per-agent policy.

Drive everything from Claude

Use the Python SDK or MCP server so an AI agent (Claude Code, Claude Desktop) can manage Kataven on your behalf.

Pick your interface

There’s no required interface — pick the one that fits.
You’re…Use
Calling the API directly from any languageREST
Writing Python (server scripts, data pipelines, CI)Python SDK
Writing Node / TypeScript on the serverNode SDK
Driving Kataven from Claude Code, Claude Desktop, Cursor chat, or any MCP-aware agentMCP Server
Embedding a voice/chat widget in your websiteBrowser SDK

What’s next

1

Get an API key

Sign in at hub.kataven.ai and grab a token.
2

Make your first call

Follow the quickstart to list your agents in 5 lines of code.
3

Build your first agent

Walk through Your first agent end-to-end.