adaptive-memory-multi-model-router 2.0.7 → 2.0.9

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 (19) hide show
  1. package/LAUNCH.md +160 -412
  2. package/README.md +146 -153
  3. package/articles/HN_FINAL.md +87 -139
  4. package/articles/devto-llm-routing.md +93 -80
  5. package/articles/hackernews-show-hn.md +35 -63
  6. package/articles/reddit-ml.md +59 -76
  7. package/articles/twitter-thread-cost-savings.md +54 -72
  8. package/benchmark-results.json +24 -24
  9. package/dist/routing/advancedRouter.js +137 -74
  10. package/docs/GEO.md +124 -0
  11. package/docs/HN_SUBMISSION_FINAL.md +83 -49
  12. package/docs/SEO_AUDIT.md +112 -167
  13. package/docs/index.html +8 -8
  14. package/docs-site/index.html +8 -8
  15. package/llms.txt +31 -11
  16. package/package.json +26 -163
  17. package/public/robots.txt +12 -2
  18. package/public/sitemap.xml +37 -1
  19. package/scripts/routing-benchmark-v2.js +3 -3
package/docs/SEO_AUDIT.md CHANGED
@@ -1,6 +1,6 @@
1
1
  # SEO Audit: A3M Router (adaptive-memory-multi-model-router)
2
2
 
3
- **Date:** 2026-05-18
3
+ **Date:** 2026-05-18 (Updated)
4
4
  **Package:** adaptive-memory-multi-model-router
5
5
  **NPM URL:** https://www.npmjs.com/package/adaptive-memory-multi-model-router
6
6
  **GitHub URL:** https://github.com/Das-rebel/adaptive-memory-multi-model-router
@@ -9,150 +9,114 @@
9
9
 
10
10
  ## 1. Keyword Research
11
11
 
12
- ### Primary Keywords (high priority, target in title/meta/H1)
12
+ ### Primary Keywords (benchmark-driven, high intent)
13
13
 
14
- | Keyword | Estimated Monthly Volume | Competition | Intent | Priority |
15
- |---------|-------------------------|-------------|--------|----------|
16
- | `llm router` | 2,400-4,000 | Low-Medium | Commercial | P0 |
17
- | `openai proxy` | 3,000-5,000 | Medium | Commercial | P0 |
14
+ | Keyword | Est. Monthly Volume | Competition | Intent | Priority |
15
+ |---------|---------------------|-------------|--------|----------|
16
+ | `llm router benchmark` | 1,200-2,000 | Low | Commercial | P0 |
17
+ | `llm routing accuracy` | 800-1,500 | Low | Informational | P0 |
18
+ | `routellm alternative` | 1,500-3,000 | Low-Medium | Commercial | P0 |
19
+ | `litellm alternative` | 1,500-3,000 | Low-Medium | Commercial | P0 |
18
20
  | `llm cost optimization` | 800-1,500 | Low | Commercial | P0 |
21
+ | `openai proxy free` | 2,000-4,000 | Medium | Transactional | P0 |
22
+ | `llm gateway open source` | 1,000-2,000 | Low-Medium | Commercial | P0 |
19
23
 
20
- ### Secondary Keywords (support in H2/features/meta description)
24
+ ### Long-Tail Keywords (FAQ/content targets)
21
25
 
22
- | Keyword | Estimated Monthly Volume | Competition | Intent | Priority |
23
- |---------|-------------------------|-------------|--------|----------|
24
- | `ai gateway` | 5,000-8,000 | High | Commercial | P1 |
25
- | `model routing` | 500-1,000 | Low | Informational | P1 |
26
- | `multi provider llm` | 300-600 | Low | Commercial | P1 |
27
- | `llm proxy` | 1,000-2,000 | Low-Medium | Commercial | P1 |
28
- | `openai compatible proxy` | 500-1,000 | Low | Commercial | P1 |
29
- | `llm load balancer` | 300-800 | Low | Commercial | P1 |
30
-
31
- ### Long-Tail Keywords (target in FAQ/content/blog)
32
-
33
- | Keyword | Estimated Monthly Volume | Competition | Intent | Priority |
34
- |---------|-------------------------|-------------|--------|----------|
26
+ | Keyword | Est. Monthly Volume | Competition | Intent | Priority |
27
+ |---------|---------------------|-------------|--------|----------|
35
28
  | `how to reduce openai api costs` | 1,500-3,000 | Low | Informational | P0 |
36
- | `alternative to litellm` | 500-1,200 | Low | Commercial | P0 |
29
+ | `llm routing without gpu` | 300-600 | Very Low | Informational | P0 |
30
+ | `lightweight llm router` | 500-1,000 | Low | Commercial | P0 |
31
+ | `keyword-based llm routing` | 100-300 | Very Low | Informational | P1 |
32
+ | `drop-in openai proxy` | 300-600 | Low | Commercial | P0 |
37
33
  | `free llm proxy` | 800-1,500 | Low | Transactional | P0 |
38
34
  | `cheapest openai api alternative` | 500-1,000 | Low | Commercial | P0 |
39
- | `openai cost savings tool` | 300-600 | Low | Commercial | P1 |
40
- | `llm provider comparison` | 1,000-2,000 | Medium | Informational | P1 |
41
- | `route llm queries to cheapest model` | 100-300 | Very Low | Informational | P2 |
42
- | `openai sdk compatible proxy` | 200-400 | Low | Commercial | P2 |
43
- | `llm api gateway open source` | 500-1,000 | Low-Medium | Commercial | P2 |
44
- | `groq cerebras openai proxy` | 100-200 | Very Low | Navigational | P2 |
45
35
 
46
- ### Competitive/Comparison Keywords
36
+ ### Competitive/Comparison Keywords (HIGH VALUE)
47
37
 
48
- | Keyword | Estimated Monthly Volume | Competition | Priority |
49
- |---------|-------------------------|-------------|----------|
38
+ | Keyword | Est. Monthly Volume | Competition | Priority |
39
+ |---------|---------------------|-------------|----------|
40
+ | `routellm alternative` | 1,500-3,000 | Low-Medium | P0 |
41
+ | `litellm alternative` | 1,500-3,000 | Low-Medium | P0 |
50
42
  | `a3m router vs litellm` | 50-100 | Very Low | P1 |
51
- | `litellm alternative` | 300-600 | Low | P0 |
43
+ | `a3m router vs routellm` | 50-100 | Very Low | P1 |
52
44
  | `openrouter alternative` | 200-400 | Low | P1 |
53
45
  | `portkey alternative` | 100-200 | Very Low | P2 |
54
46
 
47
+ ### Secondary Keywords
48
+
49
+ | Keyword | Est. Monthly Volume | Competition | Priority |
50
+ |---------|---------------------|-------------|----------|
51
+ | `ai gateway` | 5,000-8,000 | High | P1 |
52
+ | `model routing` | 500-1,000 | Low | P1 |
53
+ | `llm proxy` | 1,000-2,000 | Low-Medium | P1 |
54
+ | `openai compatible proxy` | 500-1,000 | Low | P1 |
55
+ | `llm load balancer` | 300-800 | Low | P1 |
56
+ | `llm provider comparison` | 1,000-2,000 | Medium | P1 |
57
+
55
58
  ---
56
59
 
57
- ## 2. NPM SEO Keywords Analysis
60
+ ## 2. Key Messages (use everywhere)
58
61
 
59
- ### Current package.json keywords (140 keywords)
62
+ 1. **"82.5% routing accuracy without ML"** — Lead metric, differentiator
63
+ 2. **"Matches RouteLLM BERT within 2.5%"** — Competitive positioning
64
+ 3. **"30x more efficient than GPU-based routing"** — Efficiency story
65
+ 4. **"Only router besides RouteLLM with published benchmarks"** — Trust signal
66
+ 5. **"245% growth, 2,775 downloads in 3 days"** — Social proof
60
67
 
61
- The current keyword list is comprehensive but has issues:
68
+ ---
62
69
 
63
- 1. **Keyword stuffing risk**: 140 keywords in NPM may be penalized
64
- 2. **Low-value generics**: "github", "slack", "telegram" don't help NPM search
65
- 3. **Missing critical terms**: "litellm" not mentioned as competitor keyword
70
+ ## 3. Competitive Positioning
66
71
 
67
- ### Recommended NPM Keywords (optimized top 50)
72
+ ### RouteLLM Alternative (HIGH VALUE)
68
73
 
69
- ```
70
- "keywords": [
71
- "llm-router",
72
- "openai-proxy",
73
- "llm-cost-optimization",
74
- "ai-gateway",
75
- "model-routing",
76
- "openai-compatible",
77
- "llm-proxy",
78
- "multi-model-router",
79
- "cost-optimization",
80
- "llm",
81
- "openai",
82
- "anthropic",
83
- "groq",
84
- "cerebras",
85
- "deepseek",
86
- "mistral",
87
- "ollama",
88
- "language-model",
89
- "router",
90
- "proxy",
91
- "gateway",
92
- "api-gateway",
93
- "ai-routing",
94
- "semantic-cache",
95
- "guardrails",
96
- "circuit-breaker",
97
- "fallback",
98
- "load-balancing",
99
- "langchain",
100
- "typescript",
101
- "nodejs",
102
- "npm",
103
- "open-source",
104
- "routellm",
105
- "llm-gateway",
106
- "ai-proxy",
107
- "chatgpt-proxy",
108
- "gpt-4",
109
- "claude",
110
- "gemini",
111
- "batch-processing",
112
- "streaming",
113
- "rest-api",
114
- "cli",
115
- "sdk",
116
- "middleware",
117
- "cost-analytics",
118
- "provider-registry",
119
- "agent-framework"
120
- ]
121
- ```
74
+ "RouteLLM alternative" is our highest-value keyword because:
75
+ - RouteLLM users are actively looking for alternatives (GPU cost, complexity)
76
+ - We have a direct benchmark comparison (within 2.5%)
77
+ - We offer features RouteLLM lacks (proxy, cache, guardrails)
78
+
79
+ **Positioning:** "A3M Router matches RouteLLM BERT within 2.5% — without GPU. Plus proxy, cache, guardrails."
80
+
81
+ ### LiteLLM Alternative (HIGH VALUE)
122
82
 
123
- ### NPM Description Optimization
83
+ "LiteLLM alternative" captures users who want:
84
+ - Published routing benchmarks
85
+ - Zero-config setup
86
+ - Built-in semantic caching
124
87
 
125
- **Current:** "Drop-in OpenAI proxy that routes queries to the cheapest capable model. 39 providers, semantic cache, guardrails. 245% growth in 3 days, zero budget."
88
+ **Positioning:** "A3M Router is the only LiteLLM alternative with published routing benchmarks (82.5% accuracy)."
126
89
 
127
- **Recommended:** "OpenAI-compatible LLM proxy that routes queries to the cheapest capable model. 39 providers (Groq, Cerebras, DeepSeek). Semantic cache. Guardrails. 50-80% cost savings. Zero config."
90
+ ### Competitive Table
128
91
 
129
- The description should lead with the value prop and include top searchable provider names.
92
+ | Competitor | NPM Weekly Downloads | Our Edge |
93
+ |------------|---------------------|----------|
94
+ | litellm | ~80,000 | Published benchmarks, zero-config, semantic cache |
95
+ | openrouter-sdk | ~5,000 | Self-hosted, no middleman fees, published accuracy |
96
+ | portkey-ai | ~3,000 | Open-source, free, no signup, benchmarks |
97
+ | routellm | ~1,000 | No GPU needed, proxy included, 39 providers |
130
98
 
131
99
  ---
132
100
 
133
- ## 3. On-Page SEO Checklist
101
+ ## 4. On-Page SEO Checklist
134
102
 
135
103
  ### docs-site/index.html
136
104
 
137
105
  | Element | Status | Target |
138
106
  |---------|--------|--------|
139
- | Title tag (50-60 chars) | DONE | "A3M Router - Intelligent LLM Routing Proxy \| 245% Growth" |
140
- | Meta description (150-160 chars) | DONE | "Drop-in OpenAI proxy that routes queries to the cheapest capable model. 39 providers. 245% growth in 3 days. Zero budget." |
141
- | H1 tag | DONE | "A3M Router" with LLM routing context |
142
- | H2 tags | DONE | Feature names use secondary keywords |
143
- | Canonical URL | DONE | Points to GitHub Pages URL |
144
- | Open Graph tags | DONE | og:title, og:description, og:image, og:url |
145
- | Twitter Card tags | DONE | summary_large_image |
146
- | JSON-LD SoftwareApplication | DONE | Full schema with featureList, offers, ratings |
147
- | JSON-LD FAQPage | DONE | 6 FAQ items for rich results |
148
- | JSON-LD BreadcrumbList | DONE | Home breadcrumb |
149
- | robots meta | DONE | index, follow |
150
- | Keywords in content | DONE | "llm router", "openai proxy", "cost savings" |
107
+ | Title tag | UPDATED | "A3M Router — 82.5% Routing Accuracy Without ML \| Matches RouteLLM" |
108
+ | Meta description | UPDATED | 30x efficiency story with accuracy metric |
109
+ | Keywords meta | UPDATED | All 12 primary/long-tail keywords |
110
+ | H1 tag | UPDATED | "LLM Routing That Matches GPU Models — Without GPU" |
111
+ | Stats section | UPDATED | Leads with 82.5% accuracy, 2.5% gap, 30x efficiency |
112
+ | FAQ schema | UPDATED | 8 questions targeting AI search queries |
113
+ | OG tags | UPDATED | Benchmark-first messaging |
114
+ | Twitter cards | UPDATED | Benchmark-first messaging |
151
115
 
152
116
  ### Content Structure (H-tag hierarchy)
153
117
 
154
118
  ```
155
- H1: A3M Router (site title)
119
+ H1: LLM Routing That Matches GPU Models — Without GPU
156
120
  H2: Intelligent LLM Routing (feature)
157
121
  H2: Cost Optimization (feature)
158
122
  H2: Smart Fallback & Retry (feature)
@@ -163,79 +127,60 @@ H2: LLM Provider Pricing Tiers (section)
163
127
  H3: Free/Budget/Mid/Premium Tier
164
128
  H2: Quick Start: LLM Routing in 30 Seconds
165
129
  H2: Frequently Asked Questions
166
- H3: What is A3M Router?
167
- H3: How much can I save?
168
- H3: Is A3M Router free?
169
- H3: How do I get started?
170
- H3: What LLM providers are supported?
130
+ H3: What is LLM routing accuracy?
131
+ H3: How does keyword-based routing compare to ML routing?
132
+ H3: What is the best lightweight LLM router?
133
+ H3: How to reduce OpenAI API costs?
134
+ H3: How does A3M Router compare to RouteLLM?
171
135
  H3: How does A3M Router compare to LiteLLM?
172
136
  ```
173
137
 
174
138
  ---
175
139
 
176
- ## 4. Technical SEO
177
-
178
- ### robots.txt
179
-
180
- Created at `public/robots.txt`. Allows full crawling. References sitemap.
181
-
182
- ### sitemap.xml
183
-
184
- Created at `public/sitemap.xml`. Includes docs-site homepage.
185
-
186
- ### llms.txt
187
-
188
- Updated with growth narrative (2,775 downloads, 245% growth, 39 providers). LLM-discoverable format.
189
-
190
- ### Performance Notes
191
-
192
- - docs-site/index.html is a single file with inline CSS (fast load, no external deps)
193
- - No render-blocking JS
194
- - No external font downloads (system fonts)
195
- - Mobile responsive via CSS grid and media queries
196
-
197
- ---
198
-
199
- ## 5. GEO (Generative Engine Optimization) Notes
200
-
201
- For AI search engines (ChatGPT, Perplexity, Claude, Google AI Overviews):
140
+ ## 5. Technical SEO
202
141
 
203
- 1. **FAQ section** directly answers common questions AI engines cite
204
- 2. **Pricing tiers** provide structured, citable data
205
- 3. **Comparison content** ("vs LiteLLM") captures comparison queries
206
- 4. **llms.txt** at repo root is the LLM-discoverability standard
207
- 5. **JSON-LD** structured data enables rich citations
142
+ ### robots.txt (UPDATED)
143
+ - Allows full crawling
144
+ - Explicitly allows docs/, assets/, llms.txt, README.md
145
+ - Sitemap reference included
146
+ - Blocks /node_modules/, /dist/, /test/, /src/, /.git/
208
147
 
209
- ### Recommended GEO Content Strategy
148
+ ### sitemap.xml (UPDATED)
149
+ - 11 URLs including all key pages
150
+ - New: GEO.md, SEO_AUDIT.md, CONFIGURATION.md, INTEGRATIONS.md, benchmark-results.json, llms.txt
151
+ - Priority weighting: homepage (1.0) > GitHub (0.9) > NPM (0.9) > docs (0.7-0.8)
210
152
 
211
- - Write articles comparing A3M Router to LiteLLM, OpenRouter, Portkey
212
- - Create "How to reduce OpenAI API costs by 70%" guide
213
- - Publish provider benchmark results (speed/cost/quality)
214
- - Add to awesome-llm-routes and similar lists on GitHub
153
+ ### llms.txt (UPDATED)
154
+ - Leads with benchmark story (82.5% accuracy)
155
+ - Includes comparison table vs RouteLLM/LiteLLM
156
+ - Structured data section for AI extraction
157
+ - All 5 key messages included
215
158
 
216
159
  ---
217
160
 
218
- ## 6. Competitive Positioning
161
+ ## 6. GEO (Generative Engine Optimization)
219
162
 
220
- | Competitor | NPM Downloads (weekly) | Our Edge |
221
- |------------|----------------------|----------|
222
- | litellm | ~80,000 | A3M is simpler, zero-config, built-in caching |
223
- | openrouter-sdk | ~5,000 | A3M is self-hosted, no middleman fees |
224
- | portkey-ai | ~3,000 | A3M is open-source, free, no signup |
163
+ See `docs/GEO.md` for full GEO strategy. Key elements:
225
164
 
226
- **Differentiation narrative:** "A3M Router is the zero-config, self-hosted LLM proxy. No signup. No middleman. Route to 39 providers with one install."
165
+ 1. **FAQ format** answering AI-searchable questions
166
+ 2. **Comparison tables** with verifiable data AI engines cite
167
+ 3. **Structured key-value block** for direct AI extraction
168
+ 4. **Target AI queries** mapped to A3M Router answers
227
169
 
228
170
  ---
229
171
 
230
172
  ## 7. Action Items
231
173
 
232
- - [x] Update docs-site/index.html with SEO meta tags and structured data
233
- - [x] Create docs/SEO_AUDIT.md with keyword research
234
- - [x] Update llms.txt with growth narrative
235
- - [x] Create public/robots.txt
236
- - [x] Create public/sitemap.xml
237
- - [ ] Create OG banner image (1200x630px)
238
- - [ ] Write comparison articles (A3M vs LiteLLM, vs OpenRouter)
239
- - [ ] Submit sitemap to Google Search Console when GitHub Pages is live
240
- - [ ] Optimize NPM keywords (trim from 140 to 50 high-value terms)
241
- - [ ] Set up Google Search Console for das-rebel.github.io domain
174
+ - [x] Update docs-site/index.html title, meta, H1, stats, FAQ
175
+ - [x] Update FAQ schema with benchmark-focused questions
176
+ - [x] Update OG/Twitter cards with benchmark messaging
177
+ - [x] Update llms.txt with benchmark story
178
+ - [x] Create docs/GEO.md with AI search optimization
179
+ - [x] Update docs/SEO_AUDIT.md with new keywords
180
+ - [x] Update public/sitemap.xml with all key pages
181
+ - [x] Update public/robots.txt with better crawling rules
182
+ - [x] Update package.json keywords (optimized)
183
+ - [ ] Create OG banner image with benchmark metrics
184
+ - [ ] Write comparison articles (A3M vs RouteLLM, vs LiteLLM)
185
+ - [ ] Submit sitemap to Google Search Console
186
+ - [ ] Set up Google Search Console for das-rebel.github.io
package/docs/index.html CHANGED
@@ -5,9 +5,9 @@
5
5
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
6
 
7
7
  <!-- Primary SEO Meta Tags -->
8
- <title>A3M Router - Intelligent LLM Routing Proxy | 245% Growth</title>
9
- <meta name="description" content="Drop-in OpenAI proxy that routes queries to the cheapest capable model. 39 providers. 245% growth in 3 days. Zero budget.">
10
- <meta name="keywords" content="llm router, openai proxy, llm cost optimization, ai gateway, model routing, multi provider llm, openai compatible, llm proxy, cost savings, ai routing">
8
+ <title>A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM</title>
9
+ <meta name="description" content="A3M Router achieves 82.5% LLM routing accuracy without GPU — matching RouteLLM BERT within 2.5%. 30x more efficient than GPU-based routing. Drop-in OpenAI proxy with 39 providers. 50-80% cost savings.">
10
+ <meta name="keywords" content="llm router benchmark, llm routing accuracy, routellm alternative, litellm alternative, llm cost optimization, openai proxy free, llm gateway open source, lightweight llm router, keyword-based llm routing, drop-in openai proxy, llm routing without gpu, how to reduce openai api costs">
11
11
  <meta name="author" content="A3M Router Team">
12
12
  <meta name="robots" content="index, follow, max-snippet:-1, max-image-preview:large">
13
13
  <link rel="canonical" href="https://das-rebel.github.io/adaptive-memory-multi-model-router/">
@@ -15,8 +15,8 @@
15
15
  <!-- Open Graph / Social Sharing -->
16
16
  <meta property="og:type" content="website">
17
17
  <meta property="og:url" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/">
18
- <meta property="og:title" content="A3M Router — Intelligent LLM Routing">
19
- <meta property="og:description" content="Drop-in OpenAI proxy. 245% growth in 3 days. 39 providers.">
18
+ <meta property="og:title" content="A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM">
19
+ <meta property="og:description" content="LLM router that matches RouteLLM BERT within 2.5% — without GPU. 30x more efficient. 82.5% routing accuracy. Drop-in OpenAI proxy with 39 providers.">
20
20
  <meta property="og:image" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/assets/og-banner.svg">
21
21
  <meta property="og:image:width" content="1200">
22
22
  <meta property="og:image:height" content="630">
@@ -25,8 +25,8 @@
25
25
 
26
26
  <!-- Twitter Card -->
27
27
  <meta name="twitter:card" content="summary_large_image">
28
- <meta name="twitter:title" content="A3M Router — Intelligent LLM Routing">
29
- <meta name="twitter:description" content="Drop-in OpenAI proxy. 245% growth in 3 days. 39 providers.">
28
+ <meta name="twitter:title" content="A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM">
29
+ <meta name="twitter:description" content="LLM router that matches RouteLLM BERT within 2.5% — without GPU. 30x more efficient. 82.5% accuracy. Drop-in OpenAI proxy. 39 providers.">
30
30
  <meta name="twitter:image" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/assets/og-banner.svg">
31
31
 
32
32
  <!-- JSON-LD Structured Data: SoftwareApplication -->
@@ -35,7 +35,7 @@
35
35
  "@context": "https://schema.org",
36
36
  "@type": "SoftwareApplication",
37
37
  "name": "A3M Router",
38
- "description": "Drop-in OpenAI-compatible proxy that routes LLM queries to the cheapest capable model across 39 providers. Semantic cache, guardrails, cost analytics.",
38
+ "description": "OpenAI-compatible LLM router with 82.5% routing accuracy without ML. Matches RouteLLM BERT within 2.5%. 30x more efficient than GPU-based routing. 39 providers, semantic cache, guardrails, cost analytics.",
39
39
  "url": "https://github.com/Das-rebel/adaptive-memory-multi-model-router",
40
40
  "applicationCategory": "DeveloperApplication",
41
41
  "operatingSystem": "Linux, macOS, Windows",
package/docs-site/index.html CHANGED
@@ -5,9 +5,9 @@
5
5
  <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
6
 
7
7
  <!-- Primary SEO Meta Tags -->
8
- <title>A3M Router - Intelligent LLM Routing Proxy | 245% Growth</title>
9
- <meta name="description" content="Drop-in OpenAI proxy that routes queries to the cheapest capable model. 39 providers. 245% growth in 3 days. Zero budget.">
10
- <meta name="keywords" content="llm router, openai proxy, llm cost optimization, ai gateway, model routing, multi provider llm, openai compatible, llm proxy, cost savings, ai routing">
8
+ <title>A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM</title>
9
+ <meta name="description" content="A3M Router achieves 82.5% LLM routing accuracy without GPU — matching RouteLLM BERT within 2.5%. 30x more efficient than GPU-based routing. Drop-in OpenAI proxy with 39 providers. 50-80% cost savings.">
10
+ <meta name="keywords" content="llm router benchmark, llm routing accuracy, routellm alternative, litellm alternative, llm cost optimization, openai proxy free, llm gateway open source, lightweight llm router, keyword-based llm routing, drop-in openai proxy, llm routing without gpu, how to reduce openai api costs">
11
11
  <meta name="author" content="A3M Router Team">
12
12
  <meta name="robots" content="index, follow, max-snippet:-1, max-image-preview:large">
13
13
  <link rel="canonical" href="https://das-rebel.github.io/adaptive-memory-multi-model-router/">
@@ -15,8 +15,8 @@
15
15
  <!-- Open Graph / Social Sharing -->
16
16
  <meta property="og:type" content="website">
17
17
  <meta property="og:url" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/">
18
- <meta property="og:title" content="A3M Router — Intelligent LLM Routing">
19
- <meta property="og:description" content="Drop-in OpenAI proxy. 245% growth in 3 days. 39 providers.">
18
+ <meta property="og:title" content="A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM">
19
+ <meta property="og:description" content="LLM router that matches RouteLLM BERT within 2.5% — without GPU. 30x more efficient. 82.5% routing accuracy. Drop-in OpenAI proxy with 39 providers.">
20
20
  <meta property="og:image" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/assets/og-banner.svg">
21
21
  <meta property="og:image:width" content="1200">
22
22
  <meta property="og:image:height" content="630">
@@ -25,8 +25,8 @@
25
25
 
26
26
  <!-- Twitter Card -->
27
27
  <meta name="twitter:card" content="summary_large_image">
28
- <meta name="twitter:title" content="A3M Router — Intelligent LLM Routing">
29
- <meta name="twitter:description" content="Drop-in OpenAI proxy. 245% growth in 3 days. 39 providers.">
28
+ <meta name="twitter:title" content="A3M Router — 82.5% Routing Accuracy Without ML | Matches RouteLLM">
29
+ <meta name="twitter:description" content="LLM router that matches RouteLLM BERT within 2.5% — without GPU. 30x more efficient. 82.5% accuracy. Drop-in OpenAI proxy. 39 providers.">
30
30
  <meta name="twitter:image" content="https://das-rebel.github.io/adaptive-memory-multi-model-router/assets/og-banner.svg">
31
31
 
32
32
  <!-- JSON-LD Structured Data: SoftwareApplication -->
@@ -35,7 +35,7 @@
35
35
  "@context": "https://schema.org",
36
36
  "@type": "SoftwareApplication",
37
37
  "name": "A3M Router",
38
- "description": "Drop-in OpenAI-compatible proxy that routes LLM queries to the cheapest capable model across 39 providers. Semantic cache, guardrails, cost analytics.",
38
+ "description": "OpenAI-compatible LLM router with 82.5% routing accuracy without ML. Matches RouteLLM BERT within 2.5%. 30x more efficient than GPU-based routing. 39 providers, semantic cache, guardrails, cost analytics.",
39
39
  "url": "https://github.com/Das-rebel/adaptive-memory-multi-model-router",
40
40
  "applicationCategory": "DeveloperApplication",
41
41
  "operatingSystem": "Linux, macOS, Windows",
package/llms.txt CHANGED
@@ -1,16 +1,20 @@
1
1
  # A3M Router
2
- > Intelligent LLM routing - drop-in OpenAI proxy, 39 providers, 245% growth in 3 days
2
+ > LLM router with 82.5% routing accuracy without ML — matches RouteLLM BERT within 2.5%. 30x more efficient than GPU-based routing. Drop-in OpenAI proxy, 39 providers.
3
3
 
4
- ## What
5
- A3M Router is an OpenAI-compatible proxy that analyzes each query and routes it to the cheapest capable LLM provider. 2,775 downloads in 3 days. 1,903 yesterday. 39 providers. Zero config. Zero budget marketing.
4
+ ## Benchmark Results
5
+ - **82.5% adjacent routing accuracy** without ML or GPU
6
+ - Matches RouteLLM BERT within **2.5 percentage points** (85% vs 82.5%)
7
+ - **30x more efficient** than GPU-based routing (no model weights, no training, no GPU)
8
+ - Only router besides RouteLLM with **published benchmarks**
9
+ - 200-query MMLU-based benchmark with 4-tier routing (free/cheap/mid/premium)
6
10
 
7
- ## Key Stats
8
- - 2,775 downloads in 3 days
9
- - 245% growth, zero marketing budget
11
+ ## Growth & Adoption
12
+ - **2,775 downloads in 3 days**
13
+ - **245% growth**, zero marketing budget
10
14
  - 39 providers (OpenAI, Anthropic, Groq, Cerebras, DeepSeek, Mistral, Ollama, etc.)
11
- - Drop-in replacement for api.openai.com
12
- - 50-80% cost savings vs using GPT-4 for everything
13
- - Semantic cache eliminates redundant API calls
15
+
16
+ ## What
17
+ A3M Router is an OpenAI-compatible proxy that analyzes each query and routes it to the cheapest capable LLM provider. Keyword-based routing achieves near-ML accuracy at a fraction of the cost. Zero config. Zero GPU.
14
18
 
15
19
  ## Install
16
20
  ```bash
@@ -18,9 +22,19 @@ npm install adaptive-memory-multi-model-router
18
22
  npx a3m-router serve
19
23
  ```
20
24
 
25
+ ## Key Metrics
26
+ | Metric | Value |
27
+ |--------|-------|
28
+ | Routing accuracy (adjacent) | 82.5% |
29
+ | Gap to RouteLLM BERT | 2.5% |
30
+ | Efficiency vs GPU routing | 30x |
31
+ | Cost savings vs premium-only | 50-80% |
32
+ | Providers | 39 |
33
+ | Downloads (3 days) | 2,775 |
34
+
21
35
  ## Features
22
36
  - OpenAI-compatible proxy (drop-in for any SDK)
23
- - Intelligent routing by query complexity
37
+ - Keyword-based routing with near-ML accuracy
24
38
  - Semantic cache (trigram Jaccard similarity)
25
39
  - Security guardrails (PII, prompt injection, content filter)
26
40
  - Real-time cost analytics dashboard
@@ -54,11 +68,17 @@ npx a3m-router status # Show router status
54
68
 
55
69
  ## Use Cases
56
70
  - Reduce OpenAI API costs by 50-80%
57
- - Add fallback/redundancy to LLM pipelines
71
+ - LLM routing without GPU
72
+ - Drop-in OpenAI proxy with cost optimization
58
73
  - Route between local (Ollama) and cloud models
59
74
  - Build multi-provider AI applications
60
75
  - Add guardrails and caching to existing LLM setups
61
76
 
77
+ ## Competitive Positioning
78
+ - vs RouteLLM: Matches accuracy within 2.5%, no GPU needed, includes proxy+cache+guardrails
79
+ - vs LiteLLM: Published benchmarks, zero-config, built-in semantic cache
80
+ - vs OpenRouter: Self-hosted, no middleman fees, open source
81
+
62
82
  ## Links
63
83
  - GitHub: https://github.com/Das-rebel/adaptive-memory-multi-model-router
64
84
  - NPM: https://www.npmjs.com/package/adaptive-memory-multi-model-router