The latest articles on DEV Community by voicute (@buguatec). https://dev.to/buguatec https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F4072257%2F541cd7b4-4417-4970-86b8-58dcfc9cf307.png https://dev.to/buguatec en voicute Tue, 11 Aug 2026 04:16:05 +0000 https://dev.to/buguatec/i-built-a-portable-keyword-spotting-engine-started-with-chinese-now-supporting-english-d10 https://dev.to/buguatec/i-built-a-portable-keyword-spotting-engine-started-with-chinese-now-supporting-english-d10 <h1> I built a portable keyword spotting engine — started with Chinese, now supporting English </h1> <p>I wanted offline voice commands for a few Python projects. Nothing fancy — just a single file I could drop in and run.</p> <h2> What it does </h2> <ul> <li> <strong>One Python file</strong> — <code>wakeword_engine.py</code>, copy it into your project</li> <li> <strong>Small model</strong> — ~135KB for 3 keywords, &lt;5ms inference</li> <li> <strong>Multi-keyword</strong> — one model detects 2-10 keywords in a single pass</li> <li> <strong>Fully offline</strong> — no cloud, no internet, ONNX Runtime under the hood</li> <li> <strong>5-layer noise filtering</strong> — consecutive frame check, background suppression, cooldown, burst lock, energy jump detection</li> </ul> <h2> Why Chinese first </h2> <p>Most KWS engines optimize for English. I started with Chinese instead. It's a tonal language with lots of single-syllable words — harder for keyword spotting. No pre-trained English embeddings to lean on either.</p> <p>The upside: the architecture (causal TCN on mel spectrograms) doesn't depend on any language-specific pretrained model. Once Chinese worked, English was almost free. Just needed the training data.</p> <h2> Usage </h2> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>pip <span class="nb">install </span>onnxruntime numpy pyaudio </code></pre> </div> <div class="highlight js-code-highlight"> <pre class="highlight python"><code><span class="kn">from</span> <span class="n">wakeword_engine</span> <span class="kn">import</span> <span class="n">WakeWordEngine</span> <span class="n">engine</span> <span class="o">=</span> <span class="nc">WakeWordEngine</span><span class="p">()</span> <span class="n">engine</span><span class="p">.</span><span class="nf">load</span><span class="p">(</span><span class="sh">'</span><span class="s">model_info.json</span><span class="sh">'</span><span class="p">,</span> <span class="sh">'</span><span class="s">melspectrogram.onnx</span><span class="sh">'</span><span class="p">)</span> <span class="n">engine</span><span class="p">.</span><span class="nf">set_L1</span><span class="p">(</span><span class="bp">True</span><span class="p">)</span> <span class="n">engine</span><span class="p">.</span><span class="nf">start</span><span class="p">(</span><span class="k">lambda</span> <span class="n">word</span><span class="p">,</span> <span class="n">prob</span><span class="p">,</span> <span class="n">info</span><span class="p">:</span> <span class="nf">print</span><span class="p">(</span><span class="sa">f</span><span class="sh">'</span><span class="si">{</span><span class="n">word</span><span class="si">}</span><span class="s"> </span><span class="si">{</span><span class="n">prob</span><span class="si">:</span><span class="p">.</span><span class="mi">0</span><span class="o">%</span><span class="si">}</span><span class="sh">'</span><span class="p">))</span> </code></pre> </div> <h2> What's working now </h2> <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Platform</th> <th>Status</th> </tr> </thead> <tbody> <tr> <td>Python</td> <td>✅ mature</td> </tr> <tr> <td>Web (ONNX Runtime Web)</td> <td>✅ working</td> </tr> <tr> <td>Android (Java)</td> <td>✅ working</td> </tr> <tr> <td>ESP32</td> <td>🚧 early</td> </tr> </tbody> </table></div> <h2> Next </h2> <p>Wyoming protocol integration for Home Assistant.</p> <p><strong>Repo</strong>: <a href="https://github.com/voicute/onnx-wakeword" rel="noopener noreferrer">github.com/voicute/onnx-wakeword</a> — MIT license.</p> machinelearning opensource python softwaredevelopment