@retrivora-ai/rag-engine 2.1.2 → 2.1.4

package/dist/server.mjs

@@ -576,26 +576,54 @@ function cleanApiKeyOverride(apiKeyOverride) {
   }
   return key;
 }
+async function resolveUserGatewayConfig(apiKeyOverride) {
+  var _a2;
+  if (!apiKeyOverride || !process.env.NEXT_PUBLIC_SUPABASE_URL || !process.env.SUPABASE_SERVICE_ROLE_KEY) {
+    return {};
+  }
+  try {
+    const { createAdminClient } = await import("@/lib/supabase-server");
+    const supabase = createAdminClient();
+    const { data: licenseRecord } = await supabase.from("licenses").select("project_id, customer_name, tier").eq("license_key", apiKeyOverride.trim()).single();
+    if (!licenseRecord) {
+      return {};
+    }
+    const projectId = licenseRecord.project_id;
+    const { data: configs } = await supabase.from("gateway_config").select("project_id, default_model, provider_keys, is_active").in("project_id", [projectId, "global"]).eq("is_active", true);
+    if (!configs || configs.length === 0) {
+      return {};
+    }
+    const matchedConfig = configs.find((c) => c.project_id === projectId) || configs.find((c) => c.project_id === "global");
+    const customGroqKey = (_a2 = matchedConfig == null ? void 0 : matchedConfig.provider_keys) == null ? void 0 : _a2.groq;
+    const targetModel = matchedConfig == null ? void 0 : matchedConfig.default_model;
+    return { customGroqKey, targetModel };
+  } catch (err) {
+    console.warn("[LLM Gateway Router] Error resolving user gateway_config:", err.message);
+    return {};
+  }
+}
 async function dispatchChatCompletion(req, apiKeyOverride) {
-  const effectiveKey = cleanApiKeyOverride(apiKeyOverride);
-  const provider = resolveProvider(req.model);
-  console.log(`[LLM Gateway Router] Dispatching chat request: model=${req.model}, provider=${provider}, hasGroqKey=${Boolean(process.env.GROQ_API_KEY)}, hasGeminiKey=${Boolean(process.env.GROQ_API_KEY || process.env.GEMINI_API_KEY)}, hasHFToken=${Boolean(process.env.HF_TOKEN)}`);
+  const { customGroqKey, targetModel } = await resolveUserGatewayConfig(apiKeyOverride);
+  const effectiveKey = customGroqKey || cleanApiKeyOverride(apiKeyOverride);
+  const activeModel = req.model || targetModel || "llama-3.1-8b-instant";
+  const provider = resolveProvider(activeModel);
+  console.log(`[LLM Gateway Router] Dispatching chat request: model=${activeModel}, provider=${provider}, hasCustomKey=${Boolean(customGroqKey)}, hasGlobalGroqKey=${Boolean(process.env.GROQ_API_KEY)}`);
   try {
     switch (provider) {
       case "groq":
-        return await handleGroqRequest(req, effectiveKey);
+        return await handleGroqRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }), effectiveKey);
       case "openai":
-        return await handleOpenAIRequest(req, effectiveKey);
+        return await handleOpenAIRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }), effectiveKey);
       case "gemini":
-        return await handleGeminiRequest(req, effectiveKey);
+        return await handleGeminiRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }), effectiveKey);
       case "anthropic":
-        return await handleAnthropicRequest(req, effectiveKey);
+        return await handleAnthropicRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }), effectiveKey);
       case "ollama":
-        return await handleOllamaRequest(req);
+        return await handleOllamaRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }));
       case "huggingface":
-        return await handleHuggingFaceChatRequest(req, effectiveKey);
+        return await handleHuggingFaceChatRequest(__spreadProps(__spreadValues({}, req), { model: activeModel }), effectiveKey);
       default:
-        throw new Error(`Unsupported LLM provider for model: ${req.model}`);
+        throw new Error(`Unsupported LLM provider for model: ${activeModel}`);
     }
   } catch (error) {
     console.error(`[LLM Gateway Router] Provider '${provider}' failed for model '${req.model}':`, {
@@ -3017,7 +3045,7 @@ function getRagConfig(baseConfig, env = process.env) {
   return getEnvConfig(env, baseConfig);
 }
 function getEnvConfig(env = process.env, base) {
-  var _a2, _b, _c, _d, _e, _f, _g2, _h, _i, _j, _k, _l, _m, _n, _o, _p, _q, _r, _s, _t, _u, _v, _w, _x, _y, _z, _A, _B, _C, _D, _E, _F, _G, _H, _I, _J, _K, _L, _M, _N, _O, _P, _Q, _R, _S, _T, _U, _V, _W, _X, _Y, _Z, __, _$, _aa, _ba, _ca, _da, _ea, _fa, _ga, _ha, _ia, _ja, _ka, _la, _ma, _na, _oa, _pa, _qa, _ra, _sa, _ta, _ua, _va, _wa, _xa, _ya, _za, _Aa, _Ba, _Ca, _Da, _Ea, _Fa, _Ga, _Ha, _Ia, _Ja, _Ka, _La, _Ma, _Na, _Oa, _Pa, _Qa, _Ra, _Sa, _Ta, _Ua, _Va, _Wa, _Xa, _Ya, _Za, __a, _$a, _ab, _bb, _cb, _db, _eb, _fb, _gb, _hb, _ib, _jb, _kb, _lb, _mb, _nb, _ob, _pb, _qb, _rb, _sb, _tb, _ub, _vb, _wb;
+  var _a2, _b, _c, _d, _e, _f, _g2, _h, _i, _j, _k, _l, _m, _n, _o, _p, _q, _r, _s, _t, _u, _v, _w, _x, _y, _z, _A, _B, _C, _D, _E, _F, _G, _H, _I, _J, _K, _L, _M, _N, _O, _P, _Q, _R, _S, _T, _U, _V, _W, _X, _Y, _Z, __, _$, _aa, _ba, _ca, _da, _ea, _fa, _ga, _ha, _ia, _ja, _ka, _la, _ma, _na, _oa, _pa, _qa, _ra, _sa, _ta, _ua, _va, _wa, _xa, _ya, _za, _Aa, _Ba, _Ca, _Da, _Ea, _Fa, _Ga, _Ha, _Ia, _Ja, _Ka, _La, _Ma, _Na, _Oa, _Pa, _Qa, _Ra, _Sa, _Ta, _Ua, _Va, _Wa, _Xa, _Ya, _Za, __a, _$a, _ab, _bb, _cb, _db, _eb, _fb, _gb, _hb, _ib, _jb, _kb, _lb, _mb, _nb, _ob, _pb, _qb, _rb, _sb, _tb, _ub, _vb, _wb, _xb, _yb, _zb, _Ab, _Bb, _Cb, _Db, _Eb, _Fb, _Gb, _Hb, _Ib;
   const projectId = (_c = (_b = (_a2 = readString(env, "RAG_PROJECT_ID")) != null ? _a2 : readString(env, "NEXT_PUBLIC_PROJECT_ID")) != null ? _b : base == null ? void 0 : base.projectId) != null ? _c : "__default__";
   const licenseKey = (_g2 = (_f = (_e = (_d = readString(env, "RAG_LICENSE_KEY")) != null ? _d : readString(env, "RETRIVORA_LICENSE_KEY")) != null ? _e : readString(env, "NEXT_PUBLIC_RETRIVORA_LICENSE_KEY")) != null ? _f : readString(env, "LICENSE_KEY")) != null ? _g2 : base == null ? void 0 : base.licenseKey;
   const telemetryEnabled = readString(env, "TELEMETRY_ENABLED") === "true" || readString(env, "NEXT_PUBLIC_TELEMETRY_ENABLED") === "true" || ((_h = base == null ? void 0 : base.telemetry) == null ? void 0 : _h.enabled) || Boolean(licenseKey);
@@ -3025,38 +3053,38 @@ function getEnvConfig(env = process.env, base) {
   const embeddingDimensions = readNumber(env, "EMBEDDING_DIMENSIONS", 768);
   const vectorDbOptions = {};
   if (vectorProvider === "pinecone") {
-    vectorDbOptions.apiKey = (_l = (_k = readString(env, "PINECONE_API_KEY")) != null ? _k : (_j = (_i = base == null ? void 0 : base.vectorDb) == null ? void 0 : _i.options) == null ? void 0 : _j.apiKey) != null ? _l : "";
-    vectorDbOptions.indexName = (_q = (_n = readString(env, "PINECONE_INDEX")) != null ? _n : (_m = base == null ? void 0 : base.vectorDb) == null ? void 0 : _m.indexName) != null ? _q : (_p = (_o = base == null ? void 0 : base.vectorDb) == null ? void 0 : _o.options) == null ? void 0 : _p.indexName;
+    vectorDbOptions.apiKey = (_m = (_l = (_k = readString(env, "PINECONE_API_KEY")) != null ? _k : (_j = (_i = base == null ? void 0 : base.vectorDb) == null ? void 0 : _i.options) == null ? void 0 : _j.apiKey) != null ? _l : process.env.PINECONE_API_KEY) != null ? _m : "";
+    vectorDbOptions.indexName = (_t = (_s = (_r = (_o = readString(env, "PINECONE_INDEX")) != null ? _o : (_n = base == null ? void 0 : base.vectorDb) == null ? void 0 : _n.indexName) != null ? _r : (_q = (_p = base == null ? void 0 : base.vectorDb) == null ? void 0 : _p.options) == null ? void 0 : _q.indexName) != null ? _s : process.env.PINECONE_INDEX) != null ? _t : "retrivora-free";
   } else if (vectorProvider === "pgvector" || vectorProvider === "postgresql") {
-    vectorDbOptions.connectionString = (_v = (_u = (_r = readString(env, "PGVECTOR_CONNECTION_STRING")) != null ? _r : readString(env, "POSTGRES_URL")) != null ? _u : (_t = (_s = base == null ? void 0 : base.vectorDb) == null ? void 0 : _s.options) == null ? void 0 : _t.connectionString) != null ? _v : "";
-    vectorDbOptions.tables = (_A = (_x = (_w = readString(env, "VECTOR_DB_TABLES")) != null ? _w : readString(env, "POSTGRES_TABLES")) == null ? void 0 : _x.split(",").map((t) => t.trim())) != null ? _A : (_z = (_y = base == null ? void 0 : base.vectorDb) == null ? void 0 : _y.options) == null ? void 0 : _z.tables;
-    vectorDbOptions.searchFields = (_E = (_B = readString(env, "POSTGRES_SEARCH_FIELDS")) == null ? void 0 : _B.split(",").map((f) => f.trim())) != null ? _E : (_D = (_C = base == null ? void 0 : base.vectorDb) == null ? void 0 : _C.options) == null ? void 0 : _D.searchFields;
+    vectorDbOptions.connectionString = (_y = (_x = (_u = readString(env, "PGVECTOR_CONNECTION_STRING")) != null ? _u : readString(env, "POSTGRES_URL")) != null ? _x : (_w = (_v = base == null ? void 0 : base.vectorDb) == null ? void 0 : _v.options) == null ? void 0 : _w.connectionString) != null ? _y : "";
+    vectorDbOptions.tables = (_D = (_A = (_z = readString(env, "VECTOR_DB_TABLES")) != null ? _z : readString(env, "POSTGRES_TABLES")) == null ? void 0 : _A.split(",").map((t) => t.trim())) != null ? _D : (_C = (_B = base == null ? void 0 : base.vectorDb) == null ? void 0 : _B.options) == null ? void 0 : _C.tables;
+    vectorDbOptions.searchFields = (_H = (_E = readString(env, "POSTGRES_SEARCH_FIELDS")) == null ? void 0 : _E.split(",").map((f) => f.trim())) != null ? _H : (_G = (_F = base == null ? void 0 : base.vectorDb) == null ? void 0 : _F.options) == null ? void 0 : _G.searchFields;
     vectorDbOptions.dimensions = embeddingDimensions;
   } else if (vectorProvider === "mongodb") {
-    vectorDbOptions.uri = (_I = (_H = readString(env, "MONGODB_URI")) != null ? _H : (_G = (_F = base == null ? void 0 : base.vectorDb) == null ? void 0 : _F.options) == null ? void 0 : _G.uri) != null ? _I : "";
-    vectorDbOptions.database = (_M = (_L = readString(env, "MONGODB_DB")) != null ? _L : (_K = (_J = base == null ? void 0 : base.vectorDb) == null ? void 0 : _J.options) == null ? void 0 : _K.database) != null ? _M : "";
-    vectorDbOptions.collection = (_Q = (_P = readString(env, "MONGODB_COLLECTION")) != null ? _P : (_O = (_N = base == null ? void 0 : base.vectorDb) == null ? void 0 : _N.options) == null ? void 0 : _O.collection) != null ? _Q : "";
-    vectorDbOptions.indexName = (_W = (_V = (_S = readString(env, "MONGODB_INDEX_NAME")) != null ? _S : (_R = base == null ? void 0 : base.vectorDb) == null ? void 0 : _R.indexName) != null ? _V : (_U = (_T = base == null ? void 0 : base.vectorDb) == null ? void 0 : _T.options) == null ? void 0 : _U.indexName) != null ? _W : "vector_index";
+    vectorDbOptions.uri = (_L = (_K = readString(env, "MONGODB_URI")) != null ? _K : (_J = (_I = base == null ? void 0 : base.vectorDb) == null ? void 0 : _I.options) == null ? void 0 : _J.uri) != null ? _L : "";
+    vectorDbOptions.database = (_P = (_O = readString(env, "MONGODB_DB")) != null ? _O : (_N = (_M = base == null ? void 0 : base.vectorDb) == null ? void 0 : _M.options) == null ? void 0 : _N.database) != null ? _P : "";
+    vectorDbOptions.collection = (_T = (_S = readString(env, "MONGODB_COLLECTION")) != null ? _S : (_R = (_Q = base == null ? void 0 : base.vectorDb) == null ? void 0 : _Q.options) == null ? void 0 : _R.collection) != null ? _T : "";
+    vectorDbOptions.indexName = (_Z = (_Y = (_V = readString(env, "MONGODB_INDEX_NAME")) != null ? _V : (_U = base == null ? void 0 : base.vectorDb) == null ? void 0 : _U.indexName) != null ? _Y : (_X = (_W = base == null ? void 0 : base.vectorDb) == null ? void 0 : _W.options) == null ? void 0 : _X.indexName) != null ? _Z : "vector_index";
   } else if (vectorProvider === "qdrant") {
-    vectorDbOptions.baseUrl = (__ = (_Z = readString(env, "QDRANT_URL")) != null ? _Z : (_Y = (_X = base == null ? void 0 : base.vectorDb) == null ? void 0 : _X.options) == null ? void 0 : _Y.baseUrl) != null ? __ : "http://localhost:6333";
-    vectorDbOptions.apiKey = (_ba = readString(env, "QDRANT_API_KEY")) != null ? _ba : (_aa = (_$ = base == null ? void 0 : base.vectorDb) == null ? void 0 : _$.options) == null ? void 0 : _aa.apiKey;
+    vectorDbOptions.baseUrl = (_ba = (_aa = readString(env, "QDRANT_URL")) != null ? _aa : (_$ = (__ = base == null ? void 0 : base.vectorDb) == null ? void 0 : __.options) == null ? void 0 : _$.baseUrl) != null ? _ba : "http://localhost:6333";
+    vectorDbOptions.apiKey = (_ea = readString(env, "QDRANT_API_KEY")) != null ? _ea : (_da = (_ca = base == null ? void 0 : base.vectorDb) == null ? void 0 : _ca.options) == null ? void 0 : _da.apiKey;
     vectorDbOptions.dimensions = embeddingDimensions;
   } else if (vectorProvider === "milvus") {
-    vectorDbOptions.baseUrl = (_ca = readString(env, "MILVUS_URL")) != null ? _ca : "http://localhost:19530";
+    vectorDbOptions.baseUrl = (_fa = readString(env, "MILVUS_URL")) != null ? _fa : "http://localhost:19530";
     vectorDbOptions.apiKey = readString(env, "MILVUS_API_KEY");
   } else if (vectorProvider === "chromadb") {
-    vectorDbOptions.baseUrl = (_da = readString(env, "CHROMADB_URL")) != null ? _da : "http://localhost:8000";
+    vectorDbOptions.baseUrl = (_ga = readString(env, "CHROMADB_URL")) != null ? _ga : "http://localhost:8000";
   } else if (vectorProvider === "weaviate") {
-    vectorDbOptions.baseUrl = (_ea = readString(env, "WEAVIATE_URL")) != null ? _ea : "http://localhost:8080";
+    vectorDbOptions.baseUrl = (_ha = readString(env, "WEAVIATE_URL")) != null ? _ha : "http://localhost:8080";
     vectorDbOptions.apiKey = readString(env, "WEAVIATE_API_KEY");
   } else if (vectorProvider === "redis") {
-    vectorDbOptions.baseUrl = (_fa = readString(env, "REDIS_URL")) != null ? _fa : "";
+    vectorDbOptions.baseUrl = (_ia = readString(env, "REDIS_URL")) != null ? _ia : "";
     vectorDbOptions.apiKey = readString(env, "REDIS_API_KEY");
   } else if (vectorProvider === "rest") {
-    vectorDbOptions.baseUrl = (_ga = readString(env, "VECTOR_DB_REST_URL")) != null ? _ga : "";
+    vectorDbOptions.baseUrl = (_ja = readString(env, "VECTOR_DB_REST_URL")) != null ? _ja : "";
     vectorDbOptions.headers = readString(env, "VECTOR_DB_REST_API_KEY") ? { "api-key": readString(env, "VECTOR_DB_REST_API_KEY") } : {};
   } else if (vectorProvider === "universal_rest") {
-    vectorDbOptions.baseUrl = (_ia = (_ha = readString(env, "VECTOR_BASE_URL")) != null ? _ha : readString(env, "VECTOR_DB_REST_URL")) != null ? _ia : "";
+    vectorDbOptions.baseUrl = (_la = (_ka = readString(env, "VECTOR_BASE_URL")) != null ? _ka : readString(env, "VECTOR_DB_REST_URL")) != null ? _la : "";
     vectorDbOptions.profile = readString(env, "VECTOR_UNIVERSAL_PROFILE");
     vectorDbOptions.headers = readString(env, "VECTOR_DB_REST_API_KEY") ? { Authorization: `Bearer ${readString(env, "VECTOR_DB_REST_API_KEY")}` } : {};
   }
@@ -3075,8 +3103,8 @@ function getEnvConfig(env = process.env, base) {
     // Anthropic needs a separate embedding provider; key kept for completeness
     gemini: readString(env, "GEMINI_API_KEY"),
     ollama: void 0,
-    universal_rest: (_ja = readString(env, "EMBEDDING_API_KEY")) != null ? _ja : readString(env, "OPENAI_API_KEY"),
-    custom: (_ka = readString(env, "EMBEDDING_API_KEY")) != null ? _ka : readString(env, "OPENAI_API_KEY")
+    universal_rest: (_ma = readString(env, "EMBEDDING_API_KEY")) != null ? _ma : readString(env, "OPENAI_API_KEY"),
+    custom: (_na = readString(env, "EMBEDDING_API_KEY")) != null ? _na : readString(env, "OPENAI_API_KEY")
   };
   const DEFAULT_MODEL_BY_PROVIDER = {
     openai: "gpt-4o",
@@ -3097,25 +3125,25 @@ function getEnvConfig(env = process.env, base) {
       return true;
     }
   })();
-  const DEFAULT_FREE_GATEWAY = "https://llm.retrivora.com/api/v1";
+  const defaultGatewayUrl = (_pa = (_oa = readString(env, "LITELLM_BASE_URL")) != null ? _oa : readString(env, "LLM_BASE_URL")) != null ? _pa : "https://llm.retrivora.com/api/v1";
   const defaultLlmProvider = isFreeTier ? "universal_rest" : "universal_rest";
-  const llmProvider = (_na = (_ma = readEnum(env, "LLM_PROVIDER", defaultLlmProvider, LLM_PROVIDERS)) != null ? _ma : (_la = base == null ? void 0 : base.llm) == null ? void 0 : _la.provider) != null ? _na : defaultLlmProvider;
-  const llmBaseUrl = (_qa = (_pa = readString(env, "LLM_BASE_URL")) != null ? _pa : (_oa = base == null ? void 0 : base.llm) == null ? void 0 : _oa.baseUrl) != null ? _qa : DEFAULT_FREE_GATEWAY;
-  const llmModel = (_ua = (_sa = readString(env, "LLM_MODEL")) != null ? _sa : (_ra = base == null ? void 0 : base.llm) == null ? void 0 : _ra.model) != null ? _ua : isFreeTier ? "llama-3.1-8b-instant" : (_ta = DEFAULT_MODEL_BY_PROVIDER[llmProvider]) != null ? _ta : "gpt-4o";
-  const llmApiKey = (_ya = (_xa = (_va = llmApiKeyByProvider[llmProvider]) != null ? _va : readString(env, "LLM_API_KEY")) != null ? _xa : (_wa = base == null ? void 0 : base.llm) == null ? void 0 : _wa.apiKey) != null ? _ya : "sk-retrivora-cancer-280590";
-  const llmProfile = (_Ca = (_Ba = readString(env, "LLM_UNIVERSAL_PROFILE")) != null ? _Ba : (_Aa = (_za = base == null ? void 0 : base.llm) == null ? void 0 : _za.options) == null ? void 0 : _Aa.profile) != null ? _Ca : "litellm";
+  const llmProvider = (_sa = (_ra = readEnum(env, "LLM_PROVIDER", defaultLlmProvider, LLM_PROVIDERS)) != null ? _ra : (_qa = base == null ? void 0 : base.llm) == null ? void 0 : _qa.provider) != null ? _sa : defaultLlmProvider;
+  const llmBaseUrl = (_wa = (_va = (_ta = readString(env, "LITELLM_BASE_URL")) != null ? _ta : readString(env, "LLM_BASE_URL")) != null ? _va : (_ua = base == null ? void 0 : base.llm) == null ? void 0 : _ua.baseUrl) != null ? _wa : defaultGatewayUrl;
+  const llmModel = (_Aa = (_ya = readString(env, "LLM_MODEL")) != null ? _ya : (_xa = base == null ? void 0 : base.llm) == null ? void 0 : _xa.model) != null ? _Aa : isFreeTier ? "llama-3.1-8b-instant" : (_za = DEFAULT_MODEL_BY_PROVIDER[llmProvider]) != null ? _za : "gpt-4o";
+  const llmApiKey = (_Ga = (_Fa = (_Da = (_Ca = (_Ba = llmApiKeyByProvider[llmProvider]) != null ? _Ba : readString(env, "LLM_API_KEY")) != null ? _Ca : readString(env, "LITELLM_MASTER_KEY")) != null ? _Da : readString(env, "LITELLM_API_KEY")) != null ? _Fa : (_Ea = base == null ? void 0 : base.llm) == null ? void 0 : _Ea.apiKey) != null ? _Ga : licenseKey;
+  const llmProfile = (_Ka = (_Ja = readString(env, "LLM_UNIVERSAL_PROFILE")) != null ? _Ja : (_Ia = (_Ha = base == null ? void 0 : base.llm) == null ? void 0 : _Ha.options) == null ? void 0 : _Ia.profile) != null ? _Ka : "litellm";
   const defaultEmbeddingProvider = isFreeTier ? "universal_rest" : "universal_rest";
-  const embeddingProvider = (_Fa = (_Ea = readEnum(env, "EMBEDDING_PROVIDER", defaultEmbeddingProvider, EMBEDDING_PROVIDERS)) != null ? _Ea : (_Da = base == null ? void 0 : base.embedding) == null ? void 0 : _Da.provider) != null ? _Fa : defaultEmbeddingProvider;
-  const embeddingBaseUrl = (_Ia = (_Ha = readString(env, "EMBEDDING_BASE_URL")) != null ? _Ha : (_Ga = base == null ? void 0 : base.embedding) == null ? void 0 : _Ga.baseUrl) != null ? _Ia : DEFAULT_FREE_GATEWAY;
-  const embeddingModel = (_La = (_Ka = readString(env, "EMBEDDING_MODEL")) != null ? _Ka : (_Ja = base == null ? void 0 : base.embedding) == null ? void 0 : _Ja.model) != null ? _La : "text-embedding-004";
-  const embeddingApiKey = (_Qa = (_Pa = (_Na = (_Ma = embeddingApiKeyByProvider[embeddingProvider]) != null ? _Ma : readString(env, "EMBEDDING_API_KEY")) != null ? _Na : readString(env, "LLM_API_KEY")) != null ? _Pa : (_Oa = base == null ? void 0 : base.embedding) == null ? void 0 : _Oa.apiKey) != null ? _Qa : "sk-retrivora-cancer-280590";
-  const embeddingProfile = (_Ua = (_Ta = readString(env, "EMBEDDING_UNIVERSAL_PROFILE")) != null ? _Ta : (_Sa = (_Ra = base == null ? void 0 : base.embedding) == null ? void 0 : _Ra.options) == null ? void 0 : _Sa.profile) != null ? _Ua : "litellm";
+  const embeddingProvider = (_Na = (_Ma = readEnum(env, "EMBEDDING_PROVIDER", defaultEmbeddingProvider, EMBEDDING_PROVIDERS)) != null ? _Ma : (_La = base == null ? void 0 : base.embedding) == null ? void 0 : _La.provider) != null ? _Na : defaultEmbeddingProvider;
+  const embeddingBaseUrl = (_Sa = (_Ra = (_Pa = (_Oa = readString(env, "LITELLM_BASE_URL")) != null ? _Oa : readString(env, "EMBEDDING_BASE_URL")) != null ? _Pa : readString(env, "LLM_BASE_URL")) != null ? _Ra : (_Qa = base == null ? void 0 : base.embedding) == null ? void 0 : _Qa.baseUrl) != null ? _Sa : defaultGatewayUrl;
+  const embeddingModel = (_Va = (_Ua = readString(env, "EMBEDDING_MODEL")) != null ? _Ua : (_Ta = base == null ? void 0 : base.embedding) == null ? void 0 : _Ta.model) != null ? _Va : "text-embedding-004";
+  const embeddingApiKey = (_ab = (_$a = (_Za = (_Ya = (_Xa = (_Wa = embeddingApiKeyByProvider[embeddingProvider]) != null ? _Wa : readString(env, "EMBEDDING_API_KEY")) != null ? _Xa : readString(env, "LLM_API_KEY")) != null ? _Ya : readString(env, "LITELLM_MASTER_KEY")) != null ? _Za : readString(env, "LITELLM_API_KEY")) != null ? _$a : (__a = base == null ? void 0 : base.embedding) == null ? void 0 : __a.apiKey) != null ? _ab : licenseKey;
+  const embeddingProfile = (_eb = (_db = readString(env, "EMBEDDING_UNIVERSAL_PROFILE")) != null ? _db : (_cb = (_bb = base == null ? void 0 : base.embedding) == null ? void 0 : _bb.options) == null ? void 0 : _cb.profile) != null ? _eb : "litellm";
   return __spreadProps(__spreadValues({
     projectId,
     licenseKey,
     vectorDb: {
       provider: vectorProvider,
-      indexName: (_Wa = (_Va = readString(env, "VECTOR_DB_INDEX")) != null ? _Va : vectorDbOptions.indexName) != null ? _Wa : "rag-index",
+      indexName: (_gb = (_fb = readString(env, "VECTOR_DB_INDEX")) != null ? _fb : vectorDbOptions.indexName) != null ? _gb : "rag-index",
       options: vectorDbOptions
     },
     llm: {
@@ -3145,30 +3173,30 @@ function getEnvConfig(env = process.env, base) {
       }
     },
     ui: {
-      title: (_Ya = (_Xa = readString(env, "NEXT_PUBLIC_UI_TITLE")) != null ? _Xa : readString(env, "UI_TITLE")) != null ? _Ya : "AI Assistant",
-      subtitle: (__a = (_Za = readString(env, "NEXT_PUBLIC_UI_SUBTITLE")) != null ? _Za : readString(env, "UI_SUBTITLE")) != null ? __a : "Powered by RAG",
-      primaryColor: (_ab = (_$a = readString(env, "NEXT_PUBLIC_PRIMARY_COLOR")) != null ? _$a : readString(env, "UI_PRIMARY_COLOR")) != null ? _ab : "#10b981",
-      accentColor: (_cb = (_bb = readString(env, "NEXT_PUBLIC_ACCENT_COLOR")) != null ? _bb : readString(env, "UI_ACCENT_COLOR")) != null ? _cb : "#3b82f6",
-      logoUrl: (_db = readString(env, "NEXT_PUBLIC_LOGO_URL")) != null ? _db : readString(env, "UI_LOGO_URL"),
-      placeholder: (_fb = (_eb = readString(env, "NEXT_PUBLIC_PLACEHOLDER")) != null ? _eb : readString(env, "UI_PLACEHOLDER")) != null ? _fb : "Ask me anything\u2026",
-      showSources: ((_hb = (_gb = readString(env, "NEXT_PUBLIC_SHOW_SOURCES")) != null ? _gb : readString(env, "UI_SHOW_SOURCES")) != null ? _hb : "true") !== "false",
-      welcomeMessage: (_jb = (_ib = readString(env, "NEXT_PUBLIC_WELCOME_MESSAGE")) != null ? _ib : readString(env, "UI_WELCOME_MESSAGE")) != null ? _jb : "Hello! I'm your AI assistant. Ask me anything about your documents.",
-      visualStyle: (_lb = (_kb = readString(env, "NEXT_PUBLIC_UI_VISUAL_STYLE")) != null ? _kb : readString(env, "UI_VISUAL_STYLE")) != null ? _lb : "glass",
-      borderRadius: (_nb = (_mb = readString(env, "NEXT_PUBLIC_UI_BORDER_RADIUS")) != null ? _mb : readString(env, "UI_BORDER_RADIUS")) != null ? _nb : "xl",
-      allowUpload: ((_pb = (_ob = readString(env, "NEXT_PUBLIC_ALLOW_UPLOAD")) != null ? _ob : readString(env, "UI_ALLOW_UPLOAD")) != null ? _pb : "false") === "true"
+      title: (_ib = (_hb = readString(env, "NEXT_PUBLIC_UI_TITLE")) != null ? _hb : readString(env, "UI_TITLE")) != null ? _ib : "AI Assistant",
+      subtitle: (_kb = (_jb = readString(env, "NEXT_PUBLIC_UI_SUBTITLE")) != null ? _jb : readString(env, "UI_SUBTITLE")) != null ? _kb : "Powered by RAG",
+      primaryColor: (_mb = (_lb = readString(env, "NEXT_PUBLIC_PRIMARY_COLOR")) != null ? _lb : readString(env, "UI_PRIMARY_COLOR")) != null ? _mb : "#10b981",
+      accentColor: (_ob = (_nb = readString(env, "NEXT_PUBLIC_ACCENT_COLOR")) != null ? _nb : readString(env, "UI_ACCENT_COLOR")) != null ? _ob : "#3b82f6",
+      logoUrl: (_pb = readString(env, "NEXT_PUBLIC_LOGO_URL")) != null ? _pb : readString(env, "UI_LOGO_URL"),
+      placeholder: (_rb = (_qb = readString(env, "NEXT_PUBLIC_PLACEHOLDER")) != null ? _qb : readString(env, "UI_PLACEHOLDER")) != null ? _rb : "Ask me anything\u2026",
+      showSources: ((_tb = (_sb = readString(env, "NEXT_PUBLIC_SHOW_SOURCES")) != null ? _sb : readString(env, "UI_SHOW_SOURCES")) != null ? _tb : "true") !== "false",
+      welcomeMessage: (_vb = (_ub = readString(env, "NEXT_PUBLIC_WELCOME_MESSAGE")) != null ? _ub : readString(env, "UI_WELCOME_MESSAGE")) != null ? _vb : "Hello! I'm your AI assistant. Ask me anything about your documents.",
+      visualStyle: (_xb = (_wb = readString(env, "NEXT_PUBLIC_UI_VISUAL_STYLE")) != null ? _wb : readString(env, "UI_VISUAL_STYLE")) != null ? _xb : "glass",
+      borderRadius: (_zb = (_yb = readString(env, "NEXT_PUBLIC_UI_BORDER_RADIUS")) != null ? _yb : readString(env, "UI_BORDER_RADIUS")) != null ? _zb : "xl",
+      allowUpload: ((_Bb = (_Ab = readString(env, "NEXT_PUBLIC_ALLOW_UPLOAD")) != null ? _Ab : readString(env, "UI_ALLOW_UPLOAD")) != null ? _Bb : "false") === "true"
     },
     rag: {
       topK: readNumber(env, "RAG_TOP_K", 5),
       scoreThreshold: readNumber(env, "RAG_SCORE_THRESHOLD", 0),
       chunkSize: readNumber(env, "RAG_CHUNK_SIZE", 1e3),
       chunkOverlap: readNumber(env, "RAG_CHUNK_OVERLAP", 200),
-      filterableFields: (_qb = readString(env, "RAG_FILTERABLE_FIELDS")) == null ? void 0 : _qb.split(",").map((f) => f.trim()),
+      filterableFields: (_Cb = readString(env, "RAG_FILTERABLE_FIELDS")) == null ? void 0 : _Cb.split(",").map((f) => f.trim()),
       // Query pipeline toggles — read from .env.local
       useQueryTransformation: readString(env, "RAG_USE_QUERY_TRANSFORMATION") === "true",
       useReranking: readString(env, "RAG_USE_RERANKING") === "true",
       useGraphRetrieval: readString(env, "RAG_USE_GRAPH_RETRIEVAL") === "true",
-      architecture: (_rb = readString(env, "RAG_ARCHITECTURE")) != null ? _rb : "simple",
-      chunkingStrategy: (_sb = readString(env, "RAG_CHUNKING_STRATEGY")) != null ? _sb : "recursive",
+      architecture: (_Db = readString(env, "RAG_ARCHITECTURE")) != null ? _Db : "simple",
+      chunkingStrategy: (_Eb = readString(env, "RAG_CHUNKING_STRATEGY")) != null ? _Eb : "recursive",
       uiMapping: (() => {
         const raw = readString(env, "RAG_UI_MAPPING");
         if (!raw) return void 0;
@@ -3181,7 +3209,7 @@ function getEnvConfig(env = process.env, base) {
     },
     telemetry: {
       enabled: telemetryEnabled,
-      url: (_wb = (_vb = (_tb = readString(env, "TELEMETRY_URL")) != null ? _tb : readString(env, "NEXT_PUBLIC_TELEMETRY_URL")) != null ? _vb : (_ub = base == null ? void 0 : base.telemetry) == null ? void 0 : _ub.url) != null ? _wb : process.env.NODE_ENV === "development" ? "http://localhost:3001/api/telemetry" : "https://retrivora.com/api/telemetry"
+      url: (_Ib = (_Hb = (_Fb = readString(env, "TELEMETRY_URL")) != null ? _Fb : readString(env, "NEXT_PUBLIC_TELEMETRY_URL")) != null ? _Hb : (_Gb = base == null ? void 0 : base.telemetry) == null ? void 0 : _Gb.url) != null ? _Ib : process.env.NODE_ENV === "development" ? "http://localhost:3001/api/telemetry" : "https://retrivora.com/api/telemetry"
     }
   }, readString(env, "GRAPH_DB_PROVIDER") ? {
     graphDb: {
@@ -5220,6 +5248,147 @@ var ConfigValidator = class {
   }
 };
 
+// package.json
+var package_default = {
+  name: "@retrivora-ai/rag-engine",
+  version: "2.1.4",
+  description: "Retrivora AI is a plug-and-play AI engine for RAG chat experiences \u2014 generic vector DB + LLM provider, embeddable or standalone.",
+  author: "Abhinav Alkuchi",
+  license: "UNLICENSED",
+  keywords: [
+    "rag",
+    "retrieval-augmented-generation",
+    "chatbot",
+    "vector-database",
+    "pinecone",
+    "pgvector",
+    "postgresql",
+    "milvus",
+    "qdrant",
+    "chromadb",
+    "redis",
+    "weaviate",
+    "mongodb",
+    "openai",
+    "anthropic",
+    "ollama",
+    "nextjs",
+    "react",
+    "ai",
+    "llm",
+    "embeddings"
+  ],
+  main: "dist/index.js",
+  module: "dist/index.mjs",
+  types: "dist/index.d.ts",
+  exports: {
+    ".": {
+      types: "./dist/index.d.ts",
+      require: "./dist/index.js",
+      import: "./dist/index.mjs"
+    },
+    "./style.css": "./dist/index.css",
+    "./handlers": {
+      types: "./dist/handlers/index.d.ts",
+      require: "./dist/handlers/index.js",
+      import: "./dist/handlers/index.mjs"
+    },
+    "./server": {
+      types: "./dist/server.d.ts",
+      require: "./dist/server.js",
+      import: "./dist/server.mjs"
+    }
+  },
+  browser: {
+    child_process: false,
+    fs: false,
+    net: false,
+    tls: false,
+    crypto: false,
+    os: false,
+    path: false,
+    stream: false,
+    http: false,
+    https: false,
+    zlib: false,
+    mongodb: false,
+    pg: false,
+    "pdf-parse": false,
+    mammoth: false
+  },
+  files: [
+    "dist",
+    "src",
+    ".env.example",
+    "README.md",
+    "FREE_TIER_ARCHITECTURE.md"
+  ],
+  scripts: {
+    dev: "npx @tailwindcss/cli -i src/tailwind.css -o src/index.css && tsup src/index.ts src/handlers/index.ts src/server.ts --format cjs,esm --watch --tsconfig tsconfig.build.json --external react,react-dom,next,mongodb,pg,openai,@anthropic-ai/sdk,@pinecone-database/pinecone,axios,lucide-react,mammoth,pdf-parse,xlsx,react-markdown,remark-gfm,next-themes,langchain,@langchain/core,@langchain/openai,llamaindex --inject-style",
+    build: "npm run build:pkg",
+    "build:pkg": "npx @tailwindcss/cli -i src/tailwind.css -o src/index.css && tsup src/index.ts src/handlers/index.ts src/server.ts --format cjs,esm --dts --clean --no-splitting --tsconfig tsconfig.build.json --external react,react-dom,next,mongodb,pg,openai,@anthropic-ai/sdk,@pinecone-database/pinecone,axios,lucide-react,mammoth,pdf-parse,xlsx,react-markdown,remark-gfm,next-themes,langchain,@langchain/core,@langchain/openai,llamaindex --inject-style && npx @tailwindcss/cli -i src/tailwind.css -o dist/index.css",
+    lint: "eslint",
+    clean: "rm -rf dist"
+  },
+  peerDependencies: {
+    next: ">=15.0.0",
+    react: ">=18.0.0",
+    "react-dom": ">=18.0.0"
+  },
+  peerDependenciesMeta: {
+    next: {
+      optional: true
+    }
+  },
+  dependencies: {
+    "@anthropic-ai/sdk": "^0.95.1",
+    "@google/genai": "^0.8.0",
+    "@google/generative-ai": "^0.24.1",
+    "@pinecone-database/pinecone": "^7.2.0",
+    "@types/papaparse": "^5.5.2",
+    axios: "^1.15.0",
+    "lucide-react": "^1.8.0",
+    "next-themes": "^0.4.6",
+    openai: "^6.34.0",
+    papaparse: "^5.5.3",
+    "react-is": "^18.3.1",
+    "react-markdown": "^10.1.0",
+    recharts: "^3.8.1",
+    "remark-gfm": "^4.0.1",
+    xlsx: "^0.18.5"
+  },
+  optionalDependencies: {
+    "@langchain/core": "^1.1.42",
+    "@langchain/openai": "^1.4.5",
+    langchain: "^1.3.5",
+    llamaindex: "^0.11.9",
+    mammoth: "^1.8.0",
+    mongodb: "^7.1.1",
+    "pdf-parse": "^1.1.1",
+    pg: "^8.20.0"
+  },
+  devDependencies: {
+    "@tailwindcss/cli": "^4",
+    "@tailwindcss/postcss": "^4",
+    "@types/estree": "^1.0.9",
+    "@types/node": "^20",
+    "@types/pdf-parse": "^1.1.5",
+    "@types/pg": "^8.20.0",
+    "@types/react": "^19.2.14",
+    "@types/react-dom": "^19.2.3",
+    dotenv: "^17.4.2",
+    eslint: "^9",
+    "eslint-config-next": "16.2.4",
+    next: "16.2.10",
+    tailwindcss: "^4",
+    tsup: "^8.5.1",
+    typescript: "^5"
+  }
+};
+
+// src/version.ts
+var SDK_VERSION = package_default.version || "2.1.3";
+
 // src/rag/DocumentChunker.ts
 var DocumentChunker = class {
   constructor(chunkSize = 1e3, chunkOverlap = 200, separators = ["\n# ", "\n## ", "\n### ", "\n#### ", "\n\n", "\n", " ", ""]) {
@@ -9503,6 +9672,7 @@ ${context}`;
           const telemetryUrl = ((_B = this.config.telemetry) == null ? void 0 : _B.url) || process.env.TELEMETRY_URL || process.env.NEXT_PUBLIC_TELEMETRY_URL || defaultUrl;
           const absoluteUrl = telemetryUrl.startsWith("http") ? telemetryUrl : (process.env.NEXT_PUBLIC_APP_URL || `http://localhost:${process.env.PORT || 3e3}`) + telemetryUrl;
           (async () => {
+            var _a3, _b2, _c2, _d2, _e2;
             try {
               let finalTrace = trace;
               if (!awaitHallucination && runHallucination) {
@@ -9511,6 +9681,12 @@ ${context}`;
                   finalTrace = buildTrace(backgroundScoreResult);
                 }
               }
+              const modelName = (finalTrace == null ? void 0 : finalTrace.model) || ((_a3 = this.config.llm) == null ? void 0 : _a3.model) || "llama-3.1-8b-instant";
+              const providerName = (finalTrace == null ? void 0 : finalTrace.provider) || ((_b2 = this.config.llm) == null ? void 0 : _b2.provider) || "groq";
+              const tokenCount = Number(((_c2 = finalTrace == null ? void 0 : finalTrace.tokens) == null ? void 0 : _c2.totalTokens) || 0);
+              const costEst = Number(((_d2 = finalTrace == null ? void 0 : finalTrace.tokens) == null ? void 0 : _d2.estimatedCostUsd) || 0);
+              const latencyDuration = Number(((_e2 = finalTrace == null ? void 0 : finalTrace.latency) == null ? void 0 : _e2.totalMs) || 0);
+              console.log(`[Retrivora Pipeline Telemetry] \u{1F4CA} Dispatching RAG telemetry -> Project: "${ns}", Model: "${providerName}/${modelName}", Tokens: ${tokenCount}, Latency: ${latencyDuration}ms`);
               await fetch(absoluteUrl, {
                 method: "POST",
                 headers: {
@@ -9519,7 +9695,20 @@ ${context}`;
                 body: JSON.stringify({
                   trace: finalTrace,
                   licenseKey: this.config.licenseKey,
-                  projectId: ns
+                  projectId: ns,
+                  organization: process.env.RETRIVORA_ORGANIZATION || "default-org",
+                  sdkVersion: SDK_VERSION,
+                  model: modelName,
+                  provider: providerName,
+                  tokens: tokenCount,
+                  cost: costEst,
+                  costUsd: costEst,
+                  latencyMs: latencyDuration,
+                  status: "success",
+                  action: "RAG_QUERY",
+                  feature: "RAG",
+                  timestamp: (/* @__PURE__ */ new Date()).toISOString(),
+                  details: `Org: default-org | Model: ${providerName}/${modelName} | Tokens: ${tokenCount} | Latency: ${latencyDuration}ms | Cost: $${costEst} | SDK: v${SDK_VERSION} | Feature: RAG`
                 })
               });
             } catch (err) {
@@ -10915,7 +11104,7 @@ function getOrCreatePlugin(configOrPlugin) {
   return _g[cacheKey];
 }
 function reportTelemetry(req, plugin, action, status, details, trace) {
-  var _a2;
+  var _a2, _b, _c, _d, _e, _f, _g2, _h;
   try {
     const config = plugin.getConfig();
     const licenseKey = config.licenseKey || process.env.RAG_LICENSE_KEY || process.env.RETRIVORA_LICENSE_KEY || process.env.NEXT_PUBLIC_RETRIVORA_LICENSE_KEY;
@@ -10924,29 +11113,49 @@ function reportTelemetry(req, plugin, action, status, details, trace) {
     const defaultUrl = process.env.NODE_ENV === "development" && !process.env.TELEMETRY_URL && !process.env.NEXT_PUBLIC_TELEMETRY_URL ? "http://localhost:3001/api/telemetry" : "https://retrivora.com/api/telemetry";
     const telemetryUrl = (telemetryConfig == null ? void 0 : telemetryConfig.url) || process.env.TELEMETRY_URL || process.env.NEXT_PUBLIC_TELEMETRY_URL || defaultUrl;
     const host = req.headers.get("host") || "localhost";
+    const userAgent = req.headers.get("user-agent") || `Retrivora-SDK/${SDK_VERSION}`;
     let absoluteUrl = telemetryUrl;
     if (!telemetryUrl.startsWith("http")) {
       const proto = req.headers.get("x-forwarded-proto") || "http";
       absoluteUrl = `${proto}://${host}${telemetryUrl}`;
     }
     const projectId = config.projectId || "default";
+    const model = (trace == null ? void 0 : trace.model) || ((_b = config.llm) == null ? void 0 : _b.model) || ((_c = config.embedding) == null ? void 0 : _c.model) || "llama-3.1-8b-instant";
+    const provider = (trace == null ? void 0 : trace.provider) || ((_d = config.llm) == null ? void 0 : _d.provider) || ((_e = config.embedding) == null ? void 0 : _e.provider) || "groq";
+    const tokens = Number(((_f = trace == null ? void 0 : trace.tokens) == null ? void 0 : _f.totalTokens) || (trace == null ? void 0 : trace.totalTokens) || 0);
+    const costUsd = Number(((_g2 = trace == null ? void 0 : trace.tokens) == null ? void 0 : _g2.estimatedCostUsd) || (trace == null ? void 0 : trace.costUsd) || 0);
+    const latencyMs = Number(((_h = trace == null ? void 0 : trace.latency) == null ? void 0 : _h.totalMs) || (trace == null ? void 0 : trace.latencyMs) || 0);
+    const payload = {
+      trace,
+      licenseKey,
+      projectId,
+      organization: process.env.RETRIVORA_ORGANIZATION || "default-org",
+      sdkVersion: SDK_VERSION,
+      model,
+      provider,
+      tokens,
+      cost: costUsd,
+      costUsd,
+      latencyMs,
+      status,
+      action,
+      feature: action,
+      userAgent,
+      timestamp: (/* @__PURE__ */ new Date()).toISOString(),
+      details
+    };
+    console.log(`[Retrivora SDK Telemetry] \u{1F4CA} Dispatching telemetry -> Feature: "${action}", Status: "${status}", Project: "${projectId}", Model: "${provider}/${model}", Latency: ${latencyMs}ms`);
     fetch(absoluteUrl, {
       method: "POST",
       headers: {
         "Content-Type": "application/json",
         "x-forwarded-for": req.headers.get("x-forwarded-for") || "",
-        "x-real-ip": req.headers.get("x-real-ip") || ""
+        "x-real-ip": req.headers.get("x-real-ip") || "",
+        "user-agent": userAgent
       },
-      body: JSON.stringify({
-        trace,
-        licenseKey,
-        projectId,
-        action,
-        status,
-        details
-      })
+      body: JSON.stringify(payload)
     }).catch((err) => {
-      console.warn("[Retrivora Telemetry] Async report warning:", err.message);
+      console.warn("[Retrivora SDK Telemetry] Async report warning:", err.message);
     });
   } catch (e) {
   }
@@ -11600,6 +11809,7 @@ export {
   Rule6SmallResultSetRule,
   Rule7LargeTableRule,
   RuleEngine,
+  SDK_VERSION,
   TableRendererStrategy,
   TextRendererStrategy,
   UniversalLLMAdapter,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@retrivora-ai/rag-engine",
-  "version": "2.1.2",
+  "version": "2.1.4",
   "description": "Retrivora AI is a plug-and-play AI engine for RAG chat experiences — generic vector DB + LLM provider, embeddable or standalone.",
   "author": "Abhinav Alkuchi",
   "license": "UNLICENSED",