adaptive-memory-multi-model-router 2.12.0 → 2.12.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (80) hide show
  1. package/.env +2 -0
  2. package/.github/ISSUE_TEMPLATE/bug_report.md +13 -40
  3. package/.github/ISSUE_TEMPLATE/feature_request.md +10 -27
  4. package/.github/workflows/ci.yml +8 -49
  5. package/.github/workflows/pages.yml +37 -0
  6. package/Awesome-LLM/LICENSE.md +121 -0
  7. package/Awesome-LLM/README.md +625 -0
  8. package/Awesome-LLM/contributing.md +24 -0
  9. package/Awesome-LLM/paper_list/RLHF.md +1 -0
  10. package/Awesome-LLM/paper_list/Retrieval_Augmented_Generation.md +6 -0
  11. package/Awesome-LLM/paper_list/acceleration.md +10 -0
  12. package/Awesome-LLM/paper_list/alignment.md +20 -0
  13. package/Awesome-LLM/paper_list/application.md +19 -0
  14. package/Awesome-LLM/paper_list/augmentation.md +14 -0
  15. package/Awesome-LLM/paper_list/chain_of_thougt.md +16 -0
  16. package/Awesome-LLM/paper_list/code_pretraining.md +0 -0
  17. package/Awesome-LLM/paper_list/detection.md +11 -0
  18. package/Awesome-LLM/paper_list/evaluation.md +64 -0
  19. package/Awesome-LLM/paper_list/in_context_learning.md +4 -0
  20. package/Awesome-LLM/paper_list/instruction-tuning.md +35 -0
  21. package/Awesome-LLM/paper_list/moe.md +0 -0
  22. package/Awesome-LLM/paper_list/prompt_learning.md +15 -0
  23. package/Awesome-LLM/resources/DeepSpeed_light.svg +27 -0
  24. package/Awesome-LLM/resources/alpa-logo-cropped.png +0 -0
  25. package/Awesome-LLM/resources/colossal-ai_logo_vertical.png +0 -0
  26. package/Awesome-LLM/resources/creepy_llm.jpeg +0 -0
  27. package/Awesome-LLM/resources/image8.gif +0 -0
  28. package/Awesome-LLM/resources/jax_logo_250px.png +0 -0
  29. package/Awesome-LLM/resources/logo.png +0 -0
  30. package/Awesome-LLM/resources/nvidia.png +0 -0
  31. package/Awesome-LLM/resources/tensorflow.png +0 -0
  32. package/Awesome-LLM/resources//345/244/247/345/236/213/350/257/255/350/250/200/346/250/241/345/236/213/347/232/204/350/203/275/345/212/233/345/210/206/346/236/220/344/270/216/345/272/224/347/224/250 - 30min.pdf +0 -0
  33. package/README.md +5 -4
  34. package/SUBMISSIONS.md +43 -0
  35. package/assets/chart-cost-v2.svg +91 -0
  36. package/assets/chart-cost-v3.svg +143 -0
  37. package/assets/chart-features-v2.svg +132 -0
  38. package/assets/chart-features-v3.svg +211 -0
  39. package/assets/chart-growth-v2.svg +122 -0
  40. package/assets/chart-growth-v3.svg +189 -0
  41. package/assets/cost-simple.svg +64 -0
  42. package/assets/growth-simple.svg +69 -0
  43. package/assets/hero-diagram.svg +81 -0
  44. package/assets/logo-new.svg +21 -0
  45. package/assets/social-preview-new.svg +100 -0
  46. package/assets/social-v2.svg +130 -0
  47. package/assets/social-v3.svg +212 -0
  48. package/awesome-ai-gateways/CONTRIBUTING.md +49 -0
  49. package/awesome-ai-gateways/LICENSE +3 -0
  50. package/awesome-ai-gateways/README.md +190 -0
  51. package/awesome-selfhosted/.github/ISSUE_TEMPLATE/config.yml +5 -0
  52. package/awesome-selfhosted/.github/PULL_REQUEST_TEMPLATE.md +1 -0
  53. package/awesome-selfhosted/LICENSE +376 -0
  54. package/awesome-selfhosted/README.md +2303 -0
  55. package/awesome-selfhosted/_static/awesome.png +0 -0
  56. package/awesome-selfhosted/non-free.md +377 -0
  57. package/demo-new.tape +71 -0
  58. package/demo-real.sh +198 -0
  59. package/demo-simple.tape +205 -0
  60. package/demo.html +520 -0
  61. package/demo.tape +259 -0
  62. package/dist/cli/setupWizard.d.ts +136 -0
  63. package/dist/cli/setupWizard.js +142 -165
  64. package/dist/cli/setupWizard.js.map +1 -0
  65. package/dist/cli/tui.js +152 -0
  66. package/dist/providers/providerConfig.js +28 -1
  67. package/dist/providers/providerConfig.js.map +1 -1
  68. package/dist/server/proxyServer.js +15 -0
  69. package/docs/QUICK_START.md +43 -0
  70. package/docs/USE_CASES.md +59 -0
  71. package/docs/demo-auto.html +264 -0
  72. package/docs/demo.html +416 -0
  73. package/docs/index.html +112 -599
  74. package/index.html +667 -0
  75. package/package.json +26 -169
  76. package/src/providers/providerConfig.ts +29 -1
  77. package/docs/assets/cost-comparison.svg +0 -134
  78. package/docs/assets/growth-chart-animated.svg +0 -76
  79. package/docs/assets/og-banner.svg +0 -194
  80. package/docs/assets/social-preview.svg +0 -194
package/docs/demo.html ADDED
@@ -0,0 +1,416 @@
1
+ <!DOCTYPE html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="UTF-8">
5
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
+ <title>A3M Router - One prompt in. The right model out.</title>
7
+ <style>
8
+ * { margin: 0; padding: 0; box-sizing: border-box; }
9
+
10
+ body {
11
+ background: #0d1117;
12
+ color: #e6edf3;
13
+ font-family: 'JetBrains Mono', 'Fira Code', 'SF Mono', Monaco, monospace;
14
+ min-height: 100vh;
15
+ display: flex;
16
+ flex-direction: column;
17
+ align-items: center;
18
+ justify-content: center;
19
+ padding: 20px;
20
+ }
21
+
22
+ .container { max-width: 900px; width: 100%; }
23
+
24
+ .header { text-align: center; margin-bottom: 32px; }
25
+ h1 { font-size: 2.2rem; color: #58a6ff; margin-bottom: 8px; }
26
+ .tagline { font-size: 1.2rem; color: #a371f7; margin-bottom: 16px; }
27
+ .subtitle { color: #8b949e; font-size: 0.95rem; }
28
+
29
+ .scene {
30
+ background: #161b22;
31
+ border: 1px solid #30363d;
32
+ border-radius: 12px;
33
+ padding: 24px;
34
+ margin: 16px 0;
35
+ display: none;
36
+ animation: fadeIn 0.5s ease;
37
+ }
38
+ .scene.active { display: block; }
39
+
40
+ @keyframes fadeIn {
41
+ from { opacity: 0; transform: translateY(20px); }
42
+ to { opacity: 1; transform: translateY(0); }
43
+ }
44
+
45
+ .line { margin: 6px 0; line-height: 1.6; }
46
+ .comment { color: #8b949e; }
47
+ .keyword { color: #ff7b72; }
48
+ .string { color: #a5d6ff; }
49
+
50
+ .box {
51
+ background: #21262d;
52
+ border-radius: 8px;
53
+ padding: 16px;
54
+ margin: 12px 0;
55
+ }
56
+ .box.success { border-left: 3px solid #3fb950; }
57
+ .box.info { border-left: 3px solid #58a6ff; }
58
+ .box.warning { border-left: 3px solid #d29922; }
59
+ .box.error { border-left: 3px solid #f85149; }
60
+
61
+ .success-text { color: #3fb950; }
62
+ .info-text { color: #58a6ff; }
63
+ .warning-text { color: #d29922; }
64
+ .muted { color: #8b949e; }
65
+ .money { color: #3fb950; font-size: 1.4rem; font-weight: bold; }
66
+
67
+ .providers {
68
+ display: flex;
69
+ flex-wrap: wrap;
70
+ gap: 8px;
71
+ margin: 12px 0;
72
+ }
73
+ .provider {
74
+ background: #30363d;
75
+ padding: 4px 10px;
76
+ border-radius: 4px;
77
+ font-size: 0.8rem;
78
+ color: #8b949e;
79
+ }
80
+ .provider.highlight { background: #238636; color: white; }
81
+
82
+ .comparison {
83
+ display: flex;
84
+ gap: 20px;
85
+ margin: 16px 0;
86
+ }
87
+ .col {
88
+ flex: 1;
89
+ padding: 16px;
90
+ border-radius: 8px;
91
+ text-align: center;
92
+ }
93
+ .col.bad { background: rgba(248,81,73,0.1); border: 1px solid rgba(248,81,73,0.3); }
94
+ .col.good { background: rgba(63,185,80,0.1); border: 1px solid rgba(63,185,80,0.3); }
95
+ .col .label { font-size: 0.85rem; color: #8b949e; margin-bottom: 8px; }
96
+ .col .amount { font-size: 2rem; font-weight: bold; }
97
+ .col.bad .amount { color: #f85149; }
98
+ .col.good .amount { color: #3fb950; }
99
+ .col .sub { font-size: 0.8rem; color: #8b949e; margin-top: 4px; }
100
+
101
+ .ascii {
102
+ background: #0d1117;
103
+ padding: 12px;
104
+ border-radius: 6px;
105
+ font-size: 0.8rem;
106
+ white-space: pre;
107
+ overflow-x: auto;
108
+ }
109
+
110
+ .install-box {
111
+ background: #161b22;
112
+ border: 2px solid #58a6ff;
113
+ border-radius: 12px;
114
+ padding: 24px;
115
+ text-align: center;
116
+ }
117
+ .install-cmd {
118
+ background: #0d1117;
119
+ padding: 12px 24px;
120
+ border-radius: 6px;
121
+ font-size: 1.1rem;
122
+ color: #3fb950;
123
+ margin: 12px 0;
124
+ font-family: monospace;
125
+ }
126
+
127
+ .features {
128
+ display: grid;
129
+ grid-template-columns: repeat(2, 1fr);
130
+ gap: 12px;
131
+ margin: 16px 0;
132
+ }
133
+ .feature {
134
+ background: #21262d;
135
+ padding: 12px;
136
+ border-radius: 6px;
137
+ font-size: 0.85rem;
138
+ }
139
+ .feature-name { color: #58a6ff; font-weight: bold; margin-bottom: 4px; }
140
+
141
+ .stat-row {
142
+ display: flex;
143
+ justify-content: center;
144
+ gap: 32px;
145
+ margin: 16px 0;
146
+ }
147
+ .stat { text-align: center; }
148
+ .stat-value { font-size: 1.5rem; font-weight: bold; color: #58a6ff; }
149
+ .stat-label { font-size: 0.8rem; color: #8b949e; }
150
+
151
+ .progress { display: flex; gap: 8px; justify-content: center; margin: 20px 0; }
152
+ .dot { width: 10px; height: 10px; background: #30363d; border-radius: 50%; }
153
+ .dot.done { background: #58a6ff; }
154
+
155
+ .footer {
156
+ margin-top: 24px;
157
+ color: #8b949e;
158
+ text-align: center;
159
+ font-size: 0.85rem;
160
+ }
161
+ </style>
162
+ </head>
163
+ <body>
164
+ <div class="container">
165
+ <!-- SCENE 1: One line change -->
166
+ <div class="scene active" id="s1">
167
+ <div class="header">
168
+ <h1>🔀 A3M Router</h1>
169
+ <p class="tagline">One prompt in. The right model out.</p>
170
+ <p class="subtitle">47+ providers. Budget enforcement. Semantic cache. Intelligent failover.</p>
171
+ </div>
172
+
173
+ <div class="box error">
174
+ <div class="line comment"># Your app today — every request goes to GPT-4o</div>
175
+ <div class="line keyword">const</span> response = <span class="keyword">await</span> openai.chat.completions.create({</div>
176
+ <div class="line"> model: <span class="string">"gpt-4o"</span>,</div>
177
+ <div class="line"> messages: [{ role: <span class="string">"user"</span>, content: <span class="string">"Explain quantum entanglement"</span> }]</div>
178
+ <div class="line">});</div>
179
+ </div>
180
+
181
+ <div class="box error">
182
+ <div class="line" style="color: #f85149;">❌ $0.003 per request — no fallback if OpenAI goes down</div>
183
+ </div>
184
+
185
+ <div class="box info">
186
+ <div class="line comment"># One line change — A3M routes to the cheapest capable model</div>
187
+ <div class="line keyword">const</span> response = <span class="keyword">await</span> fetch(<span class="string">"http://localhost:8787/v1/chat/completions"</span>, {</div>
188
+ <div class="line"> body: JSON.stringify({</div>
189
+ <div class="line"> model: <span class="string">"auto"</span>, <span class="comment">// ← A3M picks the best model</span></div>
190
+ <div class="line"> messages: [{ role: <span class="string">"user"</span>, content: <span class="string">"Explain quantum entanglement"</span> }]</div>
191
+ <div class="line"> })</div>
192
+ <div class="line">});</div>
193
+ </div>
194
+
195
+ <div class="box success">
196
+ <div class="line success-text">✅ Routed to: Groq (FREE)</div>
197
+ <div class="line muted"> $0.00 per request — automatic fallback chain if Groq fails</div>
198
+ </div>
199
+ </div>
200
+
201
+ <!-- SCENE 2: Providers -->
202
+ <div class="scene" id="s2">
203
+ <h2 style="color: #58a6ff; margin-bottom: 16px;">🌏 47+ Providers</h2>
204
+
205
+ <div class="box info">
206
+ <div class="line info-text">Chinese Providers (special handling)</div>
207
+ <div class="providers">
208
+ <span class="provider highlight">Groq (FREE)</span>
209
+ <span class="provider">DeepSeek</span>
210
+ <span class="provider">Kimi</span>
211
+ <span class="provider">Qwen</span>
212
+ <span class="provider">Zhipu</span>
213
+ <span class="provider">Yi</span>
214
+ <span class="provider">Stepfun</span>
215
+ <span class="provider">Moonshot</span>
216
+ <span class="provider">01.AI</span>
217
+ <span class="provider">Tencent</span>
218
+ <span class="provider">Baidu</span>
219
+ <span class="provider">Alibaba</span>
220
+ </div>
221
+ </div>
222
+
223
+ <div class="box info">
224
+ <div class="line info-text">Global Providers</div>
225
+ <div class="providers">
226
+ <span class="provider">OpenAI</span>
227
+ <span class="provider">Anthropic</span>
228
+ <span class="provider">Google</span>
229
+ <span class="provider">Mistral</span>
230
+ <span class="provider">Cohere</span>
231
+ <span class="provider">Groq</span>
232
+ <span class="provider">Perplexity</span>
233
+ <span class="provider">AWS Bedrock</span>
234
+ <span class="provider">Azure</span>
235
+ <span class="provider">Replicate</span>
236
+ <span class="provider">HuggingFace</span>
237
+ </div>
238
+ </div>
239
+
240
+ <div class="box success">
241
+ <div class="line success-text">✅ A3M routes to the cheapest healthy provider</div>
242
+ <div class="line muted"> Health checks + circuit breakers + intelligent fallback</div>
243
+ </div>
244
+ </div>
245
+
246
+ <!-- SCENE 3: Cost Savings -->
247
+ <div class="scene" id="s3">
248
+ <h2 style="color: #3fb950; margin-bottom: 16px;">💰 Save 62% on API Costs</h2>
249
+
250
+ <div class="comparison">
251
+ <div class="col bad">
252
+ <div class="label">WITHOUT A3M</div>
253
+ <div class="amount">$0.003</div>
254
+ <div class="sub">per request (GPT-4o)</div>
255
+ <div class="sub" style="margin-top: 8px;">$3.41 per 1K requests</div>
256
+ </div>
257
+ <div class="col good">
258
+ <div class="label">WITH A3M</div>
259
+ <div class="amount">$0.00</div>
260
+ <div class="sub">47% → Groq (free)</div>
261
+ <div class="sub">30% → DeepSeek ($0.14/M)</div>
262
+ <div class="sub">23% → GPT-4o-mini ($0.15/M)</div>
263
+ <div class="sub" style="margin-top: 8px;">$1.24 per 1K requests</div>
264
+ </div>
265
+ </div>
266
+
267
+ <div class="stat-row">
268
+ <div class="stat">
269
+ <div class="stat-value">62%</div>
270
+ <div class="stat-label">Cost Savings</div>
271
+ </div>
272
+ <div class="stat">
273
+ <div class="stat-value">99.5%</div>
274
+ <div class="stat-label">Routing Accuracy</div>
275
+ </div>
276
+ <div class="stat">
277
+ <div class="stat-value">&lt;1ms</div>
278
+ <div class="stat-label">Routing Latency</div>
279
+ </div>
280
+ </div>
281
+
282
+ <div class="box success">
283
+ <div class="line success-text">💰 $2,175 saved per 1M requests</div>
284
+ <div class="line muted"> At 1000 queries/day: $547 saved yearly</div>
285
+ </div>
286
+ </div>
287
+
288
+ <!-- SCENE 4: Features -->
289
+ <div class="scene" id="s4">
290
+ <h2 style="color: #58a6ff; margin-bottom: 16px;">✨ Features</h2>
291
+
292
+ <div class="features">
293
+ <div class="feature">
294
+ <div class="feature-name">🧠 Adaptive Memory</div>
295
+ <div class="muted">Learns from your usage. Updates model quality scores with every request. No retraining.</div>
296
+ </div>
297
+ <div class="feature">
298
+ <div class="feature-name">💾 Semantic Cache</div>
299
+ <div class="muted">Embedding-based lookup. 30%+ hit rate on repeated queries.</div>
300
+ </div>
301
+ <div class="feature">
302
+ <div class="feature-name">🛡️ Budget Enforcement</div>
303
+ <div class="muted">Per-user/team caps. Alerts at 50%/80%/100%. No surprises.</div>
304
+ </div>
305
+ <div class="feature">
306
+ <div class="feature-name">🔄 Intelligent Failover</div>
307
+ <div class="muted">Circuit breaker (3 fails → 60s cooldown). Automatic fallback chains.</div>
308
+ </div>
309
+ <div class="feature">
310
+ <div class="feature-name">⚡ Per-Provider Retry</div>
311
+ <div class="muted">Custom timeout per provider. Exponential backoff. 429 handling.</div>
312
+ </div>
313
+ <div class="feature">
314
+ <div class="feature-name">🎯 12-Signal Routing</div>
315
+ <div class="muted">Domain, task, structure, action verb, multi-step. Zero ML.</div>
316
+ </div>
317
+ </div>
318
+
319
+ <div class="box info">
320
+ <div class="line info-text">Zero ML. No GPU required. Starts in &lt;100ms.</div>
321
+ </div>
322
+ </div>
323
+
324
+ <!-- SCENE 5: Failover -->
325
+ <div class="scene" id="s5">
326
+ <h2 style="color: #d29922; margin-bottom: 16px;">🔄 Intelligent Failover</h2>
327
+
328
+ <div class="ascii">
329
+ [Groq] — Attempting connection...
330
+ [Groq] — ✗ FAILED — 503 Service Unavailable
331
+ [Circuit Breaker] — Tripped after 3 failures
332
+ [DeepSeek] — HEALTHY — Switching...
333
+ [DeepSeek] — ✅ Response delivered ✓
334
+ </div>
335
+
336
+ <div class="box success">
337
+ <div class="line success-text">✅ Your app never knew there was a problem</div>
338
+ <div class="line muted"> 99.9% uptime — automatic provider health scoring + fallback</div>
339
+ </div>
340
+
341
+ <div class="box warning">
342
+ <div class="line warning-text">📊 Provider Health Monitoring</div>
343
+ <div class="line muted"> Latency + error rate → health score → automatic routing</div>
344
+ </div>
345
+ </div>
346
+
347
+ <!-- SCENE 6: One Command -->
348
+ <div class="scene" id="s6">
349
+ <div class="install-box">
350
+ <div class="tagline" style="font-size: 1.3rem;">Get started in 10 seconds</div>
351
+ <div class="install-cmd">npm install -g adaptive-memory-multi-model-router</div>
352
+ <div class="muted"># Auto-detects your API keys — zero config</div>
353
+ <div class="install-cmd" style="color: #58a6ff;">npx a3m-router serve</div>
354
+ <div class="muted"># Now your app: model: "gpt-4o" → model: "auto"</div>
355
+ </div>
356
+
357
+ <div class="stat-row" style="margin-top: 24px;">
358
+ <div class="stat">
359
+ <div class="stat-value">1 line</div>
360
+ <div class="stat-label">Code change</div>
361
+ </div>
362
+ <div class="stat">
363
+ <div class="stat-value">0</div>
364
+ <div class="stat-label">API keys to manage</div>
365
+ </div>
366
+ <div class="stat">
367
+ <div class="stat-value">&lt;100ms</div>
368
+ <div class="stat-label">Startup time</div>
369
+ </div>
370
+ </div>
371
+
372
+ <div style="text-align: center; margin-top: 24px; color: #a371f7; font-size: 1.2rem;">
373
+ ✨ One prompt in. The right model out. ✨
374
+ </div>
375
+
376
+ <div class="footer" style="margin-top: 24px;">
377
+ <div>🔗 github.com/Das-rebel/adaptive-memory-multi-model-router</div>
378
+ <div>📦 npmjs.com/package/adaptive-memory-multi-model-router</div>
379
+ </div>
380
+ </div>
381
+
382
+ <div class="progress">
383
+ <div class="dot done" id="d1"></div>
384
+ <div class="dot" id="d2"></div>
385
+ <div class="dot" id="d3"></div>
386
+ <div class="dot" id="d4"></div>
387
+ <div class="dot" id="d5"></div>
388
+ <div class="dot" id="d6"></div>
389
+ </div>
390
+ </div>
391
+
392
+ <script>
393
+ // Auto-playing demo
394
+ const scenes = 6;
395
+ const delay = 10000; // 10 seconds per scene
396
+
397
+ function showScene(n) {
398
+ document.querySelectorAll('.scene').forEach(s => s.classList.remove('active'));
399
+ document.querySelectorAll('.dot').forEach(d => d.classList.remove('done'));
400
+
401
+ document.getElementById('s' + n).classList.add('active');
402
+ for (let i = 1; i <= n; i++) {
403
+ document.getElementById('d' + i).classList.add('done');
404
+ }
405
+ }
406
+
407
+ let current = 1;
408
+ showScene(current);
409
+
410
+ setInterval(() => {
411
+ current = current % scenes + 1;
412
+ showScene(current);
413
+ }, delay);
414
+ </script>
415
+ </body>
416
+ </html>