AI voice agents can answer questions, handle customer support, and make phone calls — all powered by your own knowledge base. TTS.ai makes it possible to build and deploy a voice agent without managing infrastructure.
What Is a Voice Agent?
A voice agent combines three technologies: a large language model (LLM) for understanding and generating responses, text-to-speech for natural voice output, and speech-to-text for understanding spoken input. Add a knowledge base (RAG — Retrieval Augmented Generation), and the agent can answer questions specific to your business.
Step 1: Create Your Agent
Go to Agent Manager and create a new agent. Configure:
- Name and personality — Define who the agent is and how it should respond
- System prompt — Detailed instructions for behavior, tone, and boundaries
- Voice — Choose from any TTS voice for the agent's spoken output
- Agent type — Customer support, sales, receptionist, tutor, and more
Step 2: Add a Knowledge Base
Upload documents (PDF, DOCX, TXT) to your agent's knowledge base. TTS.ai chunks the content, indexes it with full-text search, and injects relevant passages into the agent's context when answering questions. This means your agent gives accurate, source-backed answers instead of hallucinating.
Step 3: Deploy on Your Website
Embed the voice chat widget with one line of code:
<script src="https://tts.ai/widget/chat.js"
data-agent="your-agent-slug"
data-pk="pk-tts-YOUR_KEY"></script>This creates a chat bubble in the corner of your page. Visitors can type or speak to your agent, and it responds with text and voice.
Step 4: Connect a Phone Number
TTS.ai integrates with Twilio for phone-based voice agents. You can either:
- Bring your own Twilio account — Connect your existing phone numbers
- Use managed mode — We provision a number for you (50 credits/month + 10 credits/minute)
When someone calls, the agent answers with your configured voice and personality, using your knowledge base to answer questions accurately.
How RAG Works Behind the Scenes
When a user asks a question, the system:
- Converts the question to a search query
- Searches your uploaded documents using PostgreSQL full-text search
- Retrieves the most relevant passages
- Injects them into the LLM's context as reference material
- The LLM generates a response grounded in your actual content
This dramatically reduces hallucination and ensures answers are accurate to your documentation.
Use Cases
- Customer support — Answer product questions 24/7 with your FAQ and docs
- Sales — Qualify leads and answer pricing questions
- Receptionist — Handle appointment scheduling and call routing
- Education — Tutoring agents that reference course material
- Internal tools — Knowledge base search for your team
Get started at TTS.ai Agents — create your first agent in minutes.