| <!doctype html> |
| <html lang="en"> |
| <head> |
| <meta charset="utf-8" /> |
| <meta name="viewport" content="width=device-width, initial-scale=1" /> |
| <meta |
| name="description" |
| content="Ringg Parrot STT V1 is a proprietary Hindi-English code-mixed ASR system for real-time business voice applications." |
| /> |
| <title>Ringg Parrot STT V1</title> |
| <link rel="stylesheet" href="style.css" /> |
| </head> |
| <body> |
| <header class="site-header"> |
| <nav class="nav shell" aria-label="Primary navigation"> |
| <a class="brand" href="#top" aria-label="Ringg Parrot STT V1 home"> |
| <span class="brand-mark">R</span> |
| <span>Ringg Parrot STT</span> |
| </a> |
| <div class="nav-links"> |
| <a href="#benchmarks">Benchmarks</a> |
| <a href="#integration">Integration</a> |
| <a href="#access">Access</a> |
| </div> |
| <a class="nav-cta" href="https://ringg.ai/dashboard/stt" target="_blank" rel="noreferrer"> |
| Open Playground |
| </a> |
| </nav> |
|
|
| <section class="hero shell" id="top"> |
| <div class="hero-copy"> |
| <div class="tag-row"> |
| <span class="pill">Proprietary ASR</span> |
| <span>Hindi-English code-mixed speech</span> |
| </div> |
| <h1>Production-ready speech-to-text for Hindi-English voice workflows.</h1> |
| <p class="lead"> |
| Ringg Parrot STT V1 is built for real-time voice products, AI agents, |
| contact centers, and business transcription workflows that need |
| reliable Hindi, English, and code-mixed recognition. |
| </p> |
| <div class="cta-row"> |
| <a class="button primary" href="https://ringg.ai/dashboard/stt" target="_blank" rel="noreferrer"> |
| Try the Playground |
| </a> |
| <a class="button secondary" href="https://pypi.org/project/ringglabs/" target="_blank" rel="noreferrer"> |
| View Python SDK |
| </a> |
| </div> |
| <div class="proof-row" aria-label="Product highlights"> |
| <span>60-80ms streaming latency</span> |
| <span>Pipecat VAD events</span> |
| <span>Private model access</span> |
| </div> |
| </div> |
|
|
| <aside class="hero-panel" aria-label="Product summary"> |
| <div class="panel-header"> |
| <span>Streaming STT</span> |
| <strong>V1</strong> |
| </div> |
| <div class="signal-card"> |
| <span class="signal-label">Typical latency</span> |
| <strong>60-80ms</strong> |
| <p>Designed for low-latency streaming voice experiences.</p> |
| </div> |
| <div class="panel-list"> |
| <div> |
| <span>Languages</span> |
| <strong>Hindi, English, code-mix</strong> |
| </div> |
| <div> |
| <span>Integration</span> |
| <strong>SDK and Pipecat-ready events</strong> |
| </div> |
| <div> |
| <span>Access</span> |
| <strong>Hosted evaluation and commercial access</strong> |
| </div> |
| </div> |
| </aside> |
| </section> |
| </header> |
|
|
| <main class="shell"> |
| <section class="metric-strip" aria-label="Key product metrics"> |
| <article class="metric-card"> |
| <span class="metric-value">60-80ms</span> |
| <span class="metric-label">Typical streaming latency</span> |
| </article> |
| <article class="metric-card"> |
| <span class="metric-value">Hindi + English</span> |
| <span class="metric-label">Code-mixed speech support</span> |
| </article> |
| <article class="metric-card"> |
| <span class="metric-value">Proprietary</span> |
| <span class="metric-label">Private model and implementation</span> |
| </article> |
| </section> |
|
|
| <section class="section split" id="access"> |
| <div class="section-copy"> |
| <p class="section-kicker">Access</p> |
| <h2>Evaluate in the playground. Contact RinggAI for production access.</h2> |
| <p> |
| This Space provides product information for Ringg Parrot STT V1. The |
| model weights, training code, and internal implementation are not open |
| sourced. |
| </p> |
| </div> |
| <div class="info-card"> |
| <ul class="clean-list"> |
| <li>Playground access is available at ringg.ai.</li> |
| <li>Model weights are not available for download from this Space.</li> |
| <li>Production and commercial access requires RinggAI approval.</li> |
| </ul> |
| <a class="text-link" href="mailto:sales@ringg.ai">Contact sales@ringg.ai</a> |
| </div> |
| </section> |
|
|
| <section class="section split" id="integration"> |
| <div class="section-copy"> |
| <p class="section-kicker">SDK and Integration</p> |
| <h2>Integrate with voice-agent and real-time audio pipelines.</h2> |
| <p> |
| The Ringg SDK helps developers connect Ringg STT into application |
| workflows. Ringg Parrot STT V1 is highly compatible with Pipecat |
| toolkit using built-in VAD events. |
| </p> |
| </div> |
| <div class="info-card"> |
| <ul class="clean-list"> |
| <li>Python SDK is available through the ringglabs package on PyPI.</li> |
| <li>Built for low-latency streaming speech recognition.</li> |
| <li>Supports modern voice-agent orchestration patterns.</li> |
| </ul> |
| <a class="text-link" href="https://pypi.org/project/ringglabs/" target="_blank" rel="noreferrer"> |
| View ringglabs on PyPI |
| </a> |
| </div> |
| </section> |
|
|
| <section class="section" id="benchmarks"> |
| <div class="section-heading"> |
| <p class="section-kicker">Benchmarks</p> |
| <h2>WER comparison across ASR benchmark datasets.</h2> |
| <p> |
| WER stands for Word Error Rate. Lower values indicate better |
| transcription accuracy. The lowest WER in each row is highlighted. |
| </p> |
| </div> |
|
|
| <div class="benchmark-grid"> |
| <div class="table-card"> |
| <div class="table-title"> |
| <h3>Original WER</h3> |
| <span>Lower is better</span> |
| </div> |
| <div class="table-wrap"> |
| <table> |
| <thead> |
| <tr> |
| <th>Dataset</th> |
| <th>Ringg</th> |
| <th>ElevenLabs</th> |
| <th>Deepgram</th> |
| <th>Sarvam</th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td>indictts</td> |
| <td><strong>11.58</strong></td> |
| <td>16.06</td> |
| <td>13.65</td> |
| <td>15.37</td> |
| </tr> |
| <tr> |
| <td>commonvoice</td> |
| <td><strong>14.30</strong></td> |
| <td>16.59</td> |
| <td>20.04</td> |
| <td>18.21</td> |
| </tr> |
| <tr> |
| <td>fleurs</td> |
| <td>15.20</td> |
| <td><strong>11.99</strong></td> |
| <td>17.14</td> |
| <td>16.00</td> |
| </tr> |
| <tr> |
| <td>kathbath</td> |
| <td><strong>11.78</strong></td> |
| <td>13.24</td> |
| <td>15.93</td> |
| <td>17.53</td> |
| </tr> |
| <tr> |
| <td>kathbath_noisy</td> |
| <td><strong>13.09</strong></td> |
| <td>13.14</td> |
| <td>17.44</td> |
| <td>16.19</td> |
| </tr> |
| <tr> |
| <td>mucs</td> |
| <td>14.55</td> |
| <td><strong>11.69</strong></td> |
| <td>21.97</td> |
| <td>16.72</td> |
| </tr> |
| <tr> |
| <td>Overall WER</td> |
| <td>13.79</td> |
| <td><strong>13.00</strong></td> |
| <td>19.23</td> |
| <td>16.72</td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| </div> |
|
|
| <div class="table-card"> |
| <div class="table-title"> |
| <h3>Normalized WER</h3> |
| <span>Lower is better</span> |
| </div> |
| <div class="table-wrap"> |
| <table> |
| <thead> |
| <tr> |
| <th>Dataset</th> |
| <th>Ringg</th> |
| <th>ElevenLabs</th> |
| <th>Deepgram</th> |
| <th>Sarvam</th> |
| </tr> |
| </thead> |
| <tbody> |
| <tr> |
| <td>indictts</td> |
| <td><strong>3.94</strong></td> |
| <td>8.52</td> |
| <td>6.93</td> |
| <td>7.84</td> |
| </tr> |
| <tr> |
| <td>commonvoice</td> |
| <td><strong>6.37</strong></td> |
| <td>13.02</td> |
| <td>14.88</td> |
| <td>13.06</td> |
| </tr> |
| <tr> |
| <td>fleurs</td> |
| <td>9.73</td> |
| <td><strong>7.67</strong></td> |
| <td>11.35</td> |
| <td>9.54</td> |
| </tr> |
| <tr> |
| <td>kathbath</td> |
| <td><strong>7.15</strong></td> |
| <td>10.15</td> |
| <td>11.38</td> |
| <td>10.41</td> |
| </tr> |
| <tr> |
| <td>kathbath_noisy</td> |
| <td><strong>8.37</strong></td> |
| <td>10.01</td> |
| <td>12.98</td> |
| <td>11.78</td> |
| </tr> |
| <tr> |
| <td>mucs</td> |
| <td><strong>6.28</strong></td> |
| <td>6.75</td> |
| <td>12.07</td> |
| <td>7.58</td> |
| </tr> |
| <tr> |
| <td>Overall WER</td> |
| <td><strong>7.27</strong></td> |
| <td>8.94</td> |
| <td>12.36</td> |
| <td>9.76</td> |
| </tr> |
| </tbody> |
| </table> |
| </div> |
| </div> |
| </div> |
| </section> |
|
|
| <section class="section feature-grid"> |
| <article class="feature-card"> |
| <h2>Features</h2> |
| <ul class="clean-list"> |
| <li>Hindi-English code-mixed speech recognition.</li> |
| <li>Real-time streaming transcription.</li> |
| <li>File-based transcription for common audio formats.</li> |
| <li>Low-latency inference for voice products.</li> |
| </ul> |
| </article> |
| <article class="feature-card"> |
| <h2>Supported Inputs</h2> |
| <ul class="clean-list"> |
| <li>Hindi, English, and code-mixed speech.</li> |
| <li>Clear audio with minimal background noise.</li> |
| <li>16kHz or higher sample rate recommended.</li> |
| <li>WAV, MP3, FLAC, M4A, OGG, and OPUS.</li> |
| </ul> |
| </article> |
| <article class="feature-card"> |
| <h2>Use Cases</h2> |
| <ul class="clean-list"> |
| <li>Voice assistants and AI agents.</li> |
| <li>Contact center transcription.</li> |
| <li>Meeting and conversation intelligence.</li> |
| <li>Voice search, subtitling, and accessibility workflows.</li> |
| </ul> |
| </article> |
| <article class="feature-card"> |
| <h2>Limitations</h2> |
| <ul class="clean-list"> |
| <li>Accuracy may vary with noisy or low-quality audio.</li> |
| <li>Overlapping speakers and dialect variation can affect quality.</li> |
| <li>Very long files or unsupported encodings may require preprocessing.</li> |
| <li>The hosted demo may differ from production deployment settings.</li> |
| </ul> |
| </article> |
| </section> |
|
|
| <section class="section split"> |
| <div class="section-copy"> |
| <p class="section-kicker">Benchmark Dataset</p> |
| <h2>Released benchmark data and ASR transcriptions.</h2> |
| </div> |
| <div class="info-card"> |
| <p> |
| RinggAI has released the ASR Benchmarking Open-Source Dataset, which |
| includes benchmark audio/data and transcriptions generated by Ringg, |
| ElevenLabs, Deepgram, and Sarvam. |
| </p> |
| </div> |
| </section> |
|
|
| <section class="section split"> |
| <div class="section-copy"> |
| <p class="section-kicker">Privacy and Data Notice</p> |
| <h2>Review deployment terms before using sensitive data.</h2> |
| </div> |
| <div class="info-card"> |
| <p> |
| Audio handling may depend on the selected deployment, integration, and |
| commercial terms. Review RinggAI privacy terms and deployment |
| documentation before using the service with sensitive, regulated, or |
| personally identifiable data. |
| </p> |
| </div> |
| </section> |
| </main> |
|
|
| <footer class="footer shell"> |
| <div> |
| <strong>RinggAI</strong> |
| <p>Built by the RinggAI Team.</p> |
| </div> |
| <div class="footer-links"> |
| <a href="https://ringg.ai/dashboard/stt" target="_blank" rel="noreferrer">Playground</a> |
| <a href="https://pypi.org/project/ringglabs/" target="_blank" rel="noreferrer">SDK</a> |
| <a href="https://huggingface.co/RinggAI" target="_blank" rel="noreferrer">Hugging Face</a> |
| </div> |
| </footer> |
| </body> |
| </html> |
|
|
