Skip to main content

​
The Unpod Platform

Unpod is an open-source, AI-native communication platform (github.com/unpod-ai/unpod, MIT-licensed) for creating AI agents with dedicated phone numbers. Agents handle incoming calls and messages, filter communications intelligently, and deliver actionable insights - integrated with your existing business tools. You can use it two ways:

Managed Cloud

Sign up at unpod.ai and build agents in the browser. Nothing to install, nothing to operate.

Self-Hosted / On-Prem

Run the full stack on your own infrastructure - one Docker Compose command brings up everything with working defaults.
The same platform powers both: agents you build in the cloud run identically on-prem, so you can start managed and move in-house (or the reverse) without rebuilding.

​
What the platform includes

  • AI Voice Agents - conversational agents powered by LLMs, with customizable personality, knowledge, and tools
  • Agent Studio - visual no-code builder for agent behavior, prompts, and workflows
  • Multi-channel - voice calls, WhatsApp, and email through one agent interface
  • Telephony - dedicated phone numbers with SIP trunking and call routing
  • Knowledge Base - upload documents and data sources for RAG-powered responses
  • Multi-tenant workspaces - organizations, teams, RBAC, and shared Spaces
  • Call analytics - real-time dashboards, conversation logs, performance metrics
  • Workflow automation - trigger scheduling, CRM updates, and notifications from conversations
  • Desktop app - native cross-platform client built with Tauri
Under the hood it is a monorepo of a Next.js web app, a Django REST backend, FastAPI microservices, and a real-time voice engine (LiveKit + Pipecat) - see Architecture for the full picture.

​
Getting started (managed cloud)

1

Create Your Account

Sign up at unpod.ai/auth/signup. You can use your email or Google sign-in. After verifying with OTP, you will be guided to create your first voice identity.
2

Set Up Your Space

A Space is your organization’s workspace. During onboarding, you will create one with your business details, domain, and voice profile.
3

Build Your Agent

Open AI Studio and create a voice agent. Configure its identity, persona, voice, and connect it to an AI model (OpenAI, Groq, Google, Azure).
4

Connect a Phone Number

Set up telephony by creating a bridge, assigning a phone number, and configuring a voice provider.
5

Go Live

Click Publish and your agent is ready to take calls.
All you need is an email account and a modern browser.

​
Getting started (self-hosted)

git clone https://github.com/unpod-ai/unpod
cd unpod
docker compose -f docker-compose.simple.yml up -d --build
That starts the full stack in containers - frontend on :3000, API on :8000, services on :9116 - with working defaults. The Self-Hosting Quickstart covers setup options, Configuration the environment, and Architecture the moving parts.

​
Detailed guides

Account Setup

Full signup and onboarding walkthrough.

Space View

Conversations, calls, people, and analytics.

Agent Setup

Create and configure your AI voice agent.

Knowledge Base

Upload documents and FAQs for your agent.
Writing code instead? The Speech Stack gives you the same calling infrastructure as a Python SDK.