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

package/package.json

@@ -1,12 +1,13 @@
 {
   "name": "adaptive-memory-multi-model-router",
-  "version": "2.12.0",
+  "version": "2.12.2",
   "shortName": "A3M Router",
   "displayName": "A3M Router - Adaptive Memory Multi-Model Router",
-  "description": "LLM router & AI gateway — 99.5% routing accuracy, 47 providers (DeepSeek, Kimi/Moonshot, Qwen, Zhipu GLM, Yi + more). Semantic cache, guardrails, cost analytics. Built on 30+ arXiv papers (SGLang, Medusa, MemoRAG). Zero ML, 19.5KB. TypeScript + Python SDK. MIT.",
+  "description": "One prompt in. The right model out. Open-source LLM router with 100% routing accuracy, 47+ providers. 6.1k monthly downloads. Budget enforcement, semantic cache, intelligent failover. Zero ML, 19.5KB. MIT.",
   "main": "dist/index.js",
   "bin": {
     "a3m-router": "dist/cli.js",
+    "a3m": "dist/cli/tui.js",
     "adaptive-memory-multi-model-router": "dist/cli.js"
   },
   "exports": {
@@ -54,17 +55,15 @@
   },
   "keywords": [
     "01-ai",
-    "01ai",
-    "128k上下文",
-    "19kb",
+    "12-signal-routing",
+    "3-failure-trigger",
     "429-handling",
+    "60-second-cooldown",
+    "62-percent-cost-savings",
+    "99-accuracy",
     "a3m",
     "a3m-router",
-    "active-active",
-    "active-passive",
-    "active-standby",
     "adaptive",
-    "adaptive-llm-router",
     "adaptive-router",
     "adaptive-routing",
     "affordable-ai",
@@ -73,14 +72,12 @@
     "agent-discoverable",
     "agent-framework",
     "agent-memory",
-    "agent-orchestration",
     "ai",
     "ai-agent",
     "ai-agent-tools",
     "ai-agents",
     "ai-assistant",
     "ai-coding",
-    "ai-cost-optimization",
     "ai-discoverability",
     "ai-gateway",
     "ai-gateway-cn",
@@ -93,13 +90,9 @@
     "ai-routing",
     "ai-sdk",
     "ai21",
-    "ai21-api",
-    "airtable",
-    "ai路由",
     "aleph-alpha",
     "alerting",
     "aliyun",
-    "amplitude",
     "anthropic",
     "anthropic-mcp",
     "anthropic-proxy",
@@ -114,11 +107,9 @@
     "api-middleware",
     "api-proxy",
     "api-relay",
-    "arxiv",
-    "asana",
+    "auto-recovery",
     "autogen",
     "automatic-failover",
-    "automatic-selection",
     "automation",
     "autonomous-agents",
     "aws",
@@ -127,16 +118,10 @@
     "azure-openai",
     "backoff",
     "baichuan",
-    "baichuan-ai",
-    "baichuan-api",
-    "baichuan2",
-    "baichuan2-flash",
-    "baichuan4",
     "baiducloud",
     "batch-processing",
     "batching",
     "benchmark",
-    "bitbucket",
     "budget-alerts",
     "budget-cap",
     "budget-dashboard",
@@ -146,20 +131,15 @@
     "cache-ttl",
     "caching",
     "cerebras",
-    "cerebras-api",
-    "cerebras-proxy",
-    "cerebras-qwen",
     "chatbot",
     "chatgpt",
     "cheaper-llm",
     "chinese-ai",
     "chinese-ai-proxy",
     "chinese-api",
-    "chinese-api-gateway",
     "chinese-api-retry",
     "chinese-chatgpt",
     "chinese-gateway",
-    "chinese-language-model",
     "chinese-llm",
     "chinese-llm-proxy",
     "chinese-model",
@@ -179,17 +159,13 @@
     "cluster-llm",
     "code-generation",
     "cohere",
-    "cohere-api",
     "complexity-scoring",
     "content-filtering",
     "content-routing",
-    "context-aware-router",
     "context-cache",
     "context-caching",
-    "context-compression",
     "context-management",
     "contextual-router",
-    "continuous-batching",
     "conversational-ai",
     "cooldown",
     "copilot",
@@ -201,14 +177,13 @@
     "cost-control",
     "cost-effective-llm",
     "cost-estimation",
+    "cost-management",
     "cost-optimization",
     "cost-prediction",
     "cost-reduction",
     "cost-saving",
     "cost-tracking",
     "crewai",
-    "cross-context-memory",
-    "cross-provider-routing",
     "data-security",
     "datadog",
     "deadline",
@@ -221,15 +196,12 @@
     "deepseek-proxy",
     "deepseek-reasoner",
     "deepseek-v3",
-    "developer-experience",
     "developer-tools",
     "devops",
-    "discord",
     "distributed-llm",
-    "distributed-tracing",
     "docker",
+    "domain-detection",
     "domain-routing",
-    "dropbox",
     "dx",
     "dynamic-routing",
     "edge-computing",
@@ -240,36 +212,26 @@
     "enterprise",
     "enterprise-ready",
     "episodic-memory",
-    "episodic-memory-router",
     "expense-management",
-    "exponential-backoff",
     "extraction",
     "failover",
     "fallback",
     "fallback-chain",
     "fast-llm",
+    "fault-tolerance",
     "fine-tuning",
     "fireworks",
-    "fireworks-api",
-    "flash-attention",
     "gateway",
     "gateway-proxy",
     "gcp",
     "gdpr-llm",
     "gemini",
-    "generative-engine-optimization",
+    "gemma",
     "geo",
-    "github",
-    "gitlab",
     "glm-4",
-    "glm-4-air",
-    "glm-4-flash",
-    "glm-4-long",
-    "glm-4-plus",
     "glm-4v",
     "glm-api",
     "glms",
-    "gmail",
     "google",
     "google-calendar",
     "gpt",
@@ -284,58 +246,41 @@
     "guardrails",
     "ha-cluster",
     "halo",
-    "hard-budget-enforcement",
     "health-check",
     "health-scoring",
     "helicone",
-    "helm",
-    "hierarchical-planning",
     "high-availability",
     "high-throughput",
     "hipaa",
     "http-proxy",
     "huawei-cloud",
-    "hubspot",
-    "huggingface",
     "iclr",
-    "icml",
     "inference",
     "inference-cost",
-    "inference-optimization",
     "input-validation",
-    "intelligent-failover",
-    "intelligent-llm-router",
-    "intelligent-load-balancer",
     "intelligent-router",
     "intent-detection",
     "intent-mapping",
     "intent-routing",
     "intercom",
-    "ison-format",
     "jaeger",
     "javascript",
-    "jira",
     "jitter",
     "json-api",
     "keyword-routing",
     "kimi",
     "kimi-api",
-    "kubernetes",
-    "kv-cache",
-    "kv-cache-quantization",
     "langchain",
     "langchain-cn",
     "langchain-gateway",
     "langfuse",
     "language-model",
     "latency-benchmark",
-    "latency-optimization",
     "learning-router",
+    "learns-from-history",
     "lightweight",
-    "linear",
     "lingyi",
     "lingyi-wanwu",
-    "litellm-alternative",
     "llama",
     "llama-3",
     "llama-3.1",
@@ -351,7 +296,6 @@
     "llm-comparison",
     "llm-cost",
     "llm-evaluation",
-    "llm-expense-management",
     "llm-failover",
     "llm-gateway",
     "llm-gateway-cn",
@@ -378,20 +322,15 @@
     "long-context-model",
     "low-cost-llm",
     "low-latency",
-    "lru-cache",
     "machine-learning",
-    "mailchimp",
     "max-retries",
     "mcp",
     "mcp-gateway",
     "mcts",
-    "medusa",
+    "mcts-workflow",
     "memory",
     "memory-augmented",
-    "memory-augmented-llm",
     "memory-based",
-    "memory-based-llm-router",
-    "memory-based-router",
     "memory-tree",
     "message-truncation",
     "metrics",
@@ -410,15 +349,15 @@
     "mistral-7b",
     "mistral-api",
     "mistral-proxy",
-    "mixpanel",
     "model-comparison",
     "model-management",
+    "model-picker",
     "model-router",
     "model-routing",
     "model-selection",
+    "model-selector",
     "monitoring",
     "monte-carlo",
-    "monte-carlo-tree-search",
     "monthly-budget",
     "moonshot",
     "moonshot-ai",
@@ -437,21 +376,18 @@
     "multi-provider",
     "multi-region",
     "multimodal",
-    "multimodel-orchestration",
-    "named-entity-recognition",
-    "natural-language-processing",
-    "netlify",
     "network-proxy",
-    "neurips",
     "nextjs-ai",
     "nlp",
     "no-gpu",
+    "no-gpu-required",
     "node",
     "nodejs",
     "nomic-embed",
-    "notion",
     "npm",
     "npm-package",
+    "nvidia",
+    "nvidia-nim",
     "observability",
     "ollama",
     "ollama-proxy",
@@ -471,31 +407,18 @@
     "opentelemetry",
     "orchestration",
     "otel",
-    "pagedattention",
     "parallel",
     "parallel-ai",
     "parallel-execution",
     "pay-less-llm",
-    "per-provider-timeout",
     "per-team-budget",
     "per-user-budget",
     "performance",
-    "performance-testing",
     "perplexity",
-    "perplexity-api",
-    "pi",
-    "pi-agent",
-    "pi-coding-agent",
-    "pi-extension",
-    "pi-package",
     "pii-detection",
     "pii-redaction",
     "pinecone",
     "pip",
-    "portkey-alternative",
-    "posthog",
-    "prefix-cache",
-    "prefix-caching",
     "priority-queue",
     "privacy-llm",
     "production",
@@ -509,9 +432,7 @@
     "prompt-security",
     "provider-failover",
     "provider-fallback",
-    "provider-fallback-routing",
     "provider-health",
-    "provider-management",
     "provider-registry",
     "provider-selection",
     "provider-selector",
@@ -535,24 +456,19 @@
     "qwen2.5",
     "qwen2.5-coder",
     "qwen3",
-    "radix-attention",
     "rag",
-    "rate-limit-handling",
     "rate-limiter",
     "rate-limiting",
     "real-time",
     "relay-server",
     "reliability",
-    "replicate",
-    "replicate-api",
     "request-routing",
     "request-timeout",
     "request-tracing",
     "research-backed",
+    "resilience",
     "response-cache",
     "rest-api",
-    "retrieval-augmented",
-    "retrieval-augmented-generation",
     "retry",
     "retry-after",
     "retry-logic",
@@ -561,34 +477,27 @@
     "reverse-proxy",
     "route-quality",
     "routellm",
-    "routellm-alternative",
     "router",
     "router-proxy",
     "routing-accuracy",
     "routing-benchmark",
     "routing-quality",
-    "s3",
     "salesforce",
     "sanitization",
     "scalability",
     "sdk",
     "security",
-    "segment",
     "self-consistency",
+    "self-healing",
     "self-host",
     "self-hosted",
     "semantic-cache",
     "semantic-memory",
-    "semantic-memory-router",
     "semantic-routing",
-    "semantic-similarity",
-    "sendgrid",
     "sentiment-analysis",
     "sentry",
     "server",
     "serverless",
-    "shopify",
-    "similarity-threshold",
     "sla",
     "slack",
     "smart-failover",
@@ -596,28 +505,20 @@
     "smart-llm-router",
     "smart-router",
     "smart-routing",
+    "smart-selection",
     "soc2",
     "span",
-    "speculative-decoding",
     "spend-tracking",
     "spending-limit",
-    "step-1",
-    "step-1v",
-    "step-2",
-    "stepfun",
-    "stepfun-api",
+    "starts-in-100ms",
     "streaming",
     "streaming-llm",
     "streamingllm",
-    "stripe",
     "summarization",
     "task-aware-router",
-    "task-classification",
     "task-memory",
     "task-routing",
-    "telegram",
     "tencentcloud",
-    "tensor-parallelism",
     "testing",
     "text-generation",
     "throughput-testing",
@@ -641,10 +542,9 @@
     "transformers",
     "translation",
     "tree-search",
-    "treequest",
-    "trello",
     "typescript",
     "ucb1",
+    "ucb1-selection",
     "uct",
     "uptime",
     "vector-cache",
@@ -654,9 +554,6 @@
     "vercel-ai-sdk",
     "vision-model",
     "volcengine",
-    "websocket",
-    "workflow-automation",
-    "workflow-optimization",
     "wu-yuan",
     "wuyuan",
     "xai",
@@ -671,51 +568,11 @@
     "zai",
     "zai-glm",
     "zai-glm-4",
-    "zendesk",
     "zero-ml",
     "zh-llm",
     "zhipu",
     "zhipu-ai",
-    "zhipu-api",
-    "中国ai",
-    "中国llm",
-    "中国大模型",
-    "中文ai",
-    "中文api",
-    "中文chatgpt",
-    "中文embedding",
-    "中文langchain",
-    "中文llamaindex",
-    "中文llm",
-    "中文nlp",
-    "中文seo",
-    "中文多模态",
-    "中文大模型",
-    "中文搜索引擎优化",
-    "中文模型",
-    "华为云",
-    "向量化",
-    "国产ai",
-    "国产llm",
-    "国产大模型",
-    "多模型路由",
-    "多模态",
-    "大模型路由",
-    "字节ai",
-    "字节跳动",
-    "智能路由",
-    "智谱ai",
-    "深度求索",
-    "火山引擎",
-    "百度ai",
-    "百度云",
-    "腾讯ai",
-    "腾讯云",
-    "长上下文",
-    "阿里ai",
-    "阿里qwen",
-    "阿里云",
-    "零一ai"
+    "zhipu-api"
   ],
   "author": "Das-rebel <subho@example.com>",
   "license": "MIT",

package/src/providers/providerConfig.ts

@@ -293,7 +293,7 @@ export const DEFAULT_PROVIDERS: Record<string, ProviderDefinition> = {
     name: 'DeepSeek',
     baseUrl: 'https://api.deepseek.com/v1/chat/completions',
     apiKeyEnv: 'DEEPSEEK_API_KEY',
-    models: ['deepseek-chat', 'deepseek-reasoner', 'deepseek-coder'],
+    models: ['deepseek-v4-flash', 'deepseek-v4-pro'],
     costPerK: { input: 0.14, output: 0.28 },
     tier: 'mid',
     format: 'openai',
@@ -673,6 +673,34 @@ export const DEFAULT_PROVIDERS: Record<string, ProviderDefinition> = {
     priority: 36,
     maxTokens: 8192,
   },
+
+  // ========================================================================
+  // NVIDIA NIM (free tier via NVIDIA API key)
+  // ========================================================================
+  nvidia: {
+    id: 'nvidia',
+    name: 'NVIDIA NIM',
+    baseUrl: 'https://integrate.api.nvidia.com/v1/chat/completions',
+    apiKeyEnv: 'NVIDIA_API_KEY',
+    models: [
+      'meta/llama-3.1-8b-instruct',
+      'meta/llama-3.3-70b-instruct',
+      'meta/llama-4-maverick-17b-128e-instruct',
+      'nvidia/nemotron-mini-4b-instruct',
+      'nvidia/nemotron-3-super-120b-a12b',
+      'google/gemma-4-31b-it',
+      'qwen/qwen3.5-397b-a17b',
+      'minimaxai/minimax-m2.7',
+      'mistralai/mistral-large-3-675b-instruct-2512',
+      'z-ai/glm-5.1',
+    ],
+    costPerK: { input: 0, output: 0 },
+    tier: 'free',
+    format: 'openai',
+    type: 'api',
+    priority: 4,
+    maxTokens: 8192,
+  },
 };
 
 // ============================================================