File size: 11,762 Bytes
d02cb2a
9a88e93
d02cb2a
 
9a88e93
 
 
 
 
 
d02cb2a
 
 
50ae038
 
9a88e93
50ae038
 
9a88e93
 
 
 
 
 
50ae038
 
 
9a88e93
 
50ae038
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9a88e93
50ae038
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9a88e93
 
 
50ae038
 
9a88e93
 
 
 
 
 
 
 
 
50ae038
 
9a88e93
 
 
50ae038
 
 
 
 
 
 
 
 
9a88e93
50ae038
 
9a88e93
 
50ae038
9a88e93
 
 
 
 
50ae038
 
 
 
 
 
 
 
 
9a88e93
50ae038
 
9a88e93
50ae038
 
9a88e93
 
 
 
 
 
 
 
 
50ae038
9a88e93
 
50ae038
 
9a88e93
 
 
50ae038
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9a88e93
 
50ae038
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9a88e93
 
 
 
 
50ae038
9a88e93
50ae038
9a88e93
 
 
 
 
 
50ae038
9a88e93
50ae038
9a88e93
 
 
 
 
 
50ae038
9a88e93
50ae038
9a88e93
 
 
 
 
 
50ae038
9a88e93
50ae038
9a88e93
50ae038
9a88e93
 
 
 
 
 
50ae038
 
 
9a88e93
 
50ae038
9a88e93
 
 
 
 
 
 
 
50ae038
 
 
9a88e93
 
50ae038
9a88e93
 
 
 
 
 
 
 
 
 
50ae038
9a88e93
 
 
 
 
 
 
 
 
 
d02cb2a
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
<!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>