Looks Like This Service Is On Hold
I will build a real time ai voice agent with livekit and openai
About this gig
I build voice AI agents that hold real conversations in real time for cold calling, lead qualification, customer support, or interactive demos.
The stack: LiveKit for real-time audio, Whisper/STT for listening, GPT-5.1 or Claude for reasoning, and natural TTS for speaking tuned so the conversation feels fluid, not robotic.
WHAT YOU GET
- A real-time voice agent built on the LiveKit SDK
- Custom conversation flows based on your scripts and goals
- Speech-to-text + LLM + text-to-speech pipeline, latency-optimized
- Full source code and setup documentation
RELEVANT WORK
- Built a cold-calling voice agent on LiveKit (delivered as a working client demo)
- Built an AI drone assistant that converses via Whisper + TTS while tracking a human with computer vision
- Built an AI meeting assistant that joins Google Meet live and extracts customer pain points in real time
I've shipped production LLM systems since early 2023 formerly Top Rated Plus elsewhere, 5.0 rating.
Tell me your use case (sales, support, outreach, demo) and I'll propose the exact architecture before you order.
Get to know Waseem G.
AI Agents, Voice AI and LLM Apps, Python, OpenAI, Claude
- FromPakistan
- Member sinceJul 2024
Languages
Urdu, English
FAQ
Can it make actual phone calls?
Yes, via telephony integration (Premium tier). The Basic/Standard tiers run in-browser or in-app.
How natural is the voice?
I use modern low-latency TTS; you pick the voice. Most listeners can't tell within short exchanges.
Can it use my sales script or knowledge base?
Yes, Standard tier includes custom flows built on your material.
What's the latency?
Sub-second turn-taking is the target; I optimize the full STT→LLM→TTS pipeline for it. Using faster models such as Gemini Flash could increase the speed by 3x - 5x

