LiveKit Voice Auto-Troubleshooting Agent
Budget / Salary₹75,000–150,000
TypeFreelance project
LocationRemote
Posted46 minutes ago
I need a voice-controlled support agent built on LiveKit that can open a secure remote session to a user’s Windows or macOS computer, diagnose the issue in plain language, ask for a simple “yes” or “no” confirmation, and then apply the approved fix automatically. Typical problems range from slow performance, network or printer glitches, and mis-configured settings to workflows inside one of our internal business apps, so the solution must be modular enough to add new troubleshooting scripts without rewriting the core logic.
What I expect to receive
• A LiveKit powered, two-way voice interface that stays responsive even under low bandwidth.
• Remote-control functionality for both Windows and Mac with proper permission prompts and session logging.
• An intent-recognition layer (Dialogflow, Whisper, or similar—use what you know best) that maps spoken issues to a library of scripted fixes.
• An approval gate: once the agent describes the recommended action, a confirmed “yes” triggers the change; anything else cancels it.
• Clear logs and an easy way to roll back changes for safety.
• Source code, brief setup instructions, and a short recorded demo showing the agent troubleshooting at least one OS issue and one business-app scenario.
If you already have experience combining LiveKit with RPA tools, voice NLP, or remote-desktop automation, this should feel familiar. Feel free to suggest libraries or frameworks you trust; I care more about reliability, latency, and security than about specific tech choices.
What I expect to receive
• A LiveKit powered, two-way voice interface that stays responsive even under low bandwidth.
• Remote-control functionality for both Windows and Mac with proper permission prompts and session logging.
• An intent-recognition layer (Dialogflow, Whisper, or similar—use what you know best) that maps spoken issues to a library of scripted fixes.
• An approval gate: once the agent describes the recommended action, a confirmed “yes” triggers the change; anything else cancels it.
• Clear logs and an easy way to roll back changes for safety.
• Source code, brief setup instructions, and a short recorded demo showing the agent troubleshooting at least one OS issue and one business-app scenario.
If you already have experience combining LiveKit with RPA tools, voice NLP, or remote-desktop automation, this should feel familiar. Feel free to suggest libraries or frameworks you trust; I care more about reliability, latency, and security than about specific tech choices.
Apply on Freelancer →
Project sourced from Freelancer.com. Applications happen directly on the original platform — we never collect your data.