From URL to live voice agent in 10 minutes
A complete walkthrough of the Auralius setup flow: create your agent, upload your knowledge base, connect a phone number, and take the first real call — all without writing a line of code.
You do not need to write a single line of code to deploy an Auralius voice agent. This guide walks through the complete setup — from a blank workspace to a live phone number answering real calls from your own knowledge — in under ten minutes.
Step 1 — Create your agent
In the dashboard, go to Agents and click New agent. Give it a name, choose a voice, and write a short persona description — for example:
"You are Elena, the friendly front-desk assistant at Apex Dental. You help callers book appointments, answer pricing questions, and explain insurance coverage."
That persona description becomes the agent's identity on every call. Keep it concise — one or two sentences is enough. The agent's behavior will be shaped mainly by your knowledge base in the next step.
Step 2 — Build your knowledge base
Go to Knowledge and create a new knowledge base. You can upload content three ways:
- Paste a URL. Auralius scrapes the page, chunks it, and embeds it automatically.
- Upload a PDF or document. Price lists, FAQs, policies, service menus — anything you would hand to a new employee.
- Type or paste text directly. Useful for hours of operation, custom scripts, or quick facts.
The agent does not guess or make things up. When a caller asks a question, the agent retrieves the most relevant chunk from your knowledge base and answers from that — so accuracy is bounded by what you upload.
Once created, go back to your agent and attach the knowledge base under Settings. You can link multiple knowledge bases to one agent.
Step 3 — Connect a phone number
In Phone Numbers, provision a new number. Auralius sets up the Twilio voice webhook automatically — you do not need to touch Twilio's console. Every workspace includes at least one number, even on the Free plan, allowing you to connect and test immediately.
Once provisioned, route the number to your agent from the Canvas view by drawing a connection from the phone node to the agent node. The routing is live immediately.
Step 4 — Test before going live
Use Test chat on the agent page to try questions in text before a real call. You will see which knowledge chunks the agent retrieved and how it composed the answer. This is the fastest way to spot gaps in your knowledge base.
When text looks good, place a real call to your provisioned number. The agent greets the caller, listens, and responds — including barge-in, so callers can interrupt naturally without waiting for the agent to finish speaking.
Step 5 — Review and refine
Every call is saved in Call Logs with a full turn-by-turn transcript. Read through the first few calls: wherever the agent stumbled, add or correct the relevant knowledge. Most teams see quality converge within five to ten calls.
On Pro and Scale plans, post-call automation runs after every call — generating a structured summary and triggering any webhooks you configure. You can route that summary to your CRM, calendar, or Slack channel without writing integration code.
What each plan includes
- Free — 1 agent, 20 KB documents, 50 K tokens/month, 1 BYOC phone number.
- Starter ($49/mo) — 3 agents, 200 documents, 500 K tokens, 1 phone number, post-call automation.
- Pro — 20 agents, 2 K documents, 5 M tokens, 5 phone numbers, omnichannel (WhatsApp + live chat).
- Scale — unlimited everything, custom voice, unlimited phone numbers.
You are live
That is it. No infrastructure to manage, no telephony configuration, no model fine-tuning. Your phone line now has an AI agent that answers from your own documents, handles interruptions naturally, and logs every conversation for review. Start on Free to explore the dashboard and test with a real number, then upgrade to Starter as your call volume grows.
Published August 10, 2026 · Kolerr Lab Engineering Team