Qwen Audio Agent + Claude Code: This Basically makes Claude Code a HUMAN!

In this video, I'll connect Qwen Audio Agent to Claude Code so I can keep a voice conversation going while the coding work runs in the background. First I'll explain how the voice front end, the local gateway and the Claude Code backend fit together, then install it with npm, set up the DashScope voice key, the Claude backend and the workspace, and start the gateway and the browser interface. Then I'll build a small issue tracker called VoiceDesk by voice, ask questions while it builds, interrupt the spoken answer, change a requirement mid-build and report a whitespace-title bug. Finally, I'll test cancelling a backend task, go over where this setup is useful and what it costs, and share the checks I'd judge it by. -- Key Takeaways: 🎙️ Qwen Audio Agent is an open-source app that connects a voice conversation to agents like Claude Code, with browser, terminal and desktop clients. 🔀 The voice front end handles the conversation, Claude Code does the coding, and a local gateway connects them through an ACP adapter. 💻 Setup is npm install -g qwen-audio-agent@2.0.1, a DashScope key in config.env, qwenaudio install claude, and qwenaudio setup --backend claude. ⚠️ Opening a terminal in a folder doesn't set the agent's workspace, so set the workspace path in the config and check it. 🧠 You can ask questions while Claude keeps building, but it doesn't make Claude write code any faster. ⚡ Interrupting the spoken answer stops the voice, not a task the backend has already accepted. 🔍 A spoken "it's done" isn't proof, so open the app, check the diff and run the tests. 🛑 Cancelling a task is different from muting the mic, and cancelling doesn't undo edits that were already made. 💸 The software is open source, but the DashScope voice model and your Claude usage are both billed separately. ✅ It's great for talking through features and describing bugs while the agent works; keep the keyboard for exact paths, code and errors. -- Timestamps: 0:00 Intro 0:50 What Qwen Audio Agent is 1:53 Install 2:29 Voice config: DashScope key + realtime model 3:08 Connect Claude Code as the coding backend 4:03 Set the project workspace 4:24 Start the gateway + WebUI 4:45 First voice test + Claude connection check 5:22 Building VoiceDesk by voice 6:26 Asking questions while Claude codes 7:06 Interrupting the spoken answer 7:42 Changing a requirement mid-build 8:26 Testing the finished app 9:17 Debugging: the spaces-only title bug 10:46 Cancelling a backend task 11:40 Where it's useful + what it costs 12:56 Troubleshooting tips 13:34 Final verdict