from hubui_brain import Brain, QueryContext
from openai import AsyncOpenAI
brain = Brain()
client = AsyncOpenAI()
@brain.on_query()
async def handle(ctx: QueryContext):
response = await ctx.auto_process(
client.chat.completions.create(
model="gpt-5.4",
reasoning_effort="high",
web_search_options={},
messages=[{"role": "user", "content": ctx.query}]
)
)
await ctx.reply(response.choices[0].message.content)
brain.run(port=8080)Already have an AI agent? Plug it into HubUI and give it the power to talk to humans through voice, phone, and chat.
Sign up, pick a voice, write instructions. Your agent is live on web voice and chat right away. Add phone instantly*.
pip install hubui-brain - plug in your Python backend for custom reasoning, database lookups, and API calls. Your agent gets smarter.
Your agent is already live on web voice and chat. Add a phone number* or embed with npm install @hubui/client - same agent, same Brain, available on every channel your users prefer.
Connect your AI agent to all the ways humans communicate. Voice, phone, and chat - all from one integration.
Connect your custom Python backend via WebSocket. Use OpenAI, Gemini, Anthropic, or any LLM for deep reasoning. Your Agent handles conversation autonomously; the Brain is called on-demand.
Provision phone numbers instantly. Assign them to agents, configure concurrent call channels, and start taking inbound calls.
7 premium HD voices with distinct personalities. Adjust speed from 0.5× to 1.5×. Configure silence detection, interruptions, and conversation behavior.
WebRTC-powered voice with intelligent voice activity detection. No awkward delays - conversations feel natural and human.
Embed voice and text agents in your web app with @hubui/client. WebRTC audio, microphone management, and live transcripts - all handled.
Session metrics, Brain performance tracking, call logs with searchable transcripts. Monitor latency, error rates, and usage in real-time.
HubUI separates the conversation layer from your backend logic. The HubUI agent handles the user experience. Your agent (the brain) handles the thinking. They run independently.
Real problems. Real businesses. The kind of conversations only a reasoning-backed voice agent can handle.
A clinic, salon, or service business. Patients call at 10pm to book or reschedule. The Brain checks your real scheduling system, finds available slots, and confirms the booking - staff arrive to a full calendar.
A customer calls their bank: "I'm self-employed, 680 credit score - what are my real refinancing options?" The Brain runs GPT-5.4 or Claude, reasons through the full scenario, and gives a real answer. Not a script. Not a dead end.
A company built an HR agent - it handles PTO balances, benefits questions, expense policy. It works in chat. Now employees dial in and ask "How many vacation days do I have left?" The same agent answers, over voice. Nothing in the backend changes.
Start with $200 first-adopter credit. Usage draws from your prepaid balance.
Prepaid credit for early builders launching real agents
Simple usage-based pricing after your credit is applied
For larger teams, custom usage needs, and production rollouts
Plug in your existing AI agent. Start with $200 first-adopter credit, then pay as you go for voice, AI responses, and phone numbers.