One agent. Built your way.
Build natural voice agents with knowledge and actions through the visual builder or API. Test and deploy the same saved agent to your website, phone or application, powered by Omni. The builder remains Live Beta.
- Price
- $0.08/min
- Endpoint
POST /v1/agents + wss://api.pyai.com/v1/omni- Scope
omni:session- Model id
pyai-omni-realtime
See it built, not described
Describe the job in the console, connect docs and tools, and point a number at it. The lifecycle player below walks the whole path.
Start in the consoleStart in minutes
// Agents uses the same Omni realtime engine.
const ws = new WebSocket(
"wss://api.pyai.com/v1/omni?session_label=front_desk&format=pcm16&rate=24000",
["pyai.v1", "pyai-key." + apiKey],
);Build your way
From instructions to a working voice agent
Create and manage the same agent through the console or API. Add knowledge and actions, test it, then connect your website, phone or application.
From prompt to production
Describe the job
No code · no engineeringWrite what the agent should do in plain English - the persona, the goal, the guardrails. No flow charts, no code.
No phone number required
Your website gets a Call Now button
The agent you just built can answer visitors on your own site. Configure the launcher in the console, publish, and paste one script tag: no API key on the page, no backend to run.
The whole integration
<script src="https://cdn.pyai.com/widget/v11/pyai-widget.js"
data-widget="wdgt_your_published_id" async></script>- 1.Describe the job in the console wizard and test the agent in your browser.
- 2.Set the launcher style, colors, and the sites allowed to show it.
- 3.Publish and paste the tag. The button answers with your agent's voice, knowledge, and tools.
What your visitors see
The launcher opens a live voice session with the same agent you tested in the console. Pill, orb, card, or inline, in your colors and your label.
Language support
Agents across every supported language
Use one PyAI product surface across the language your customers speak.
English
English
Explore Agents
Español
Spanish
Explore Agents
Français
French
Explore Agents
हिन्दी
Hindi
Explore Agents
Deutsch
German
Explore Agents
Italiano
Italian
Explore Agents
Português
Portuguese
Explore Agents
日本語
Japanese
Explore Agents
한국어
Korean
Explore Agents
Kreyòl ayisyen
Haitian Creole
Explore Agents
普通话
Mandarin Chinese
Explore Agents
العربية
Arabic
Explore Agents
What you get
Create in the UI or API
Describe the job in the builder or create a saved profile with the API. Both use the same agent record, knowledge bindings and actions.
Evals before volume
Test the conversations that matter before rollout, then monitor real calls as the agent scales.
Recap-ready
Add Recap to turn every Agent call into summaries, dispositions, next steps, and searchable call intelligence.
Telephony as a separate line item
Use managed telephony alongside the agent when you want PyAI to handle the phone path.
Capabilities
- Visual builder + Agents API
- Realtime playground
- Knowledge bases + tools
- Website, phone + application deployment
- Recap add-on
- Trace add-on
Built on Omni's realtime engine.
FAQ
Are Omni and Agents separate products?
Agents is the workspace for building and running voice agents. Omni is its realtime runtime. Create a saved agent through the console or POST /v1/agents, then connect with its id as the session_label on /v1/omni. Existing Omni integrations continue to work.
Can I use the API without creating a saved agent?
Yes. Connect directly to the Omni realtime endpoint and send your configuration for each session. Saved profiles are optional, and agents created with a project API key appear in that same project in the console.
Can I put the agent on my website?
Yes. Publish any agent as a Call Now button: pick the launcher style in the console, publish, and paste one script tag into your site. Visitors talk to the same agent by voice from the page, with its knowledge and tools. It needs no API key on the page, no backend, and no phone number.
What does it cost?
Agents is Live Beta and usage is billed at the current Agents rate. Managed telephony is metered separately. Check the pricing page for current rates before production use.
AtomsAI implementation services
Have our team build it with you
AtomsAI, our AI services business, provides white-glove implementation for your voice agents, from bespoke setup to ongoing improvement.
Bespoke setup
Turn your business requirements into agent instructions, knowledge and workflows.
Custom integrations
Connect your agent to your CRM, internal tools and deployment environment.
Ongoing improvement
Refine conversations, build evaluations and improve your agent as your needs evolve.
Optional services, scoped and quoted separately from PyAI usage.
Launch an Agent in the console.
Build and test through the console or API, then fund live usage when you are ready for production calls. The builder is Live Beta.