@blockrun/llm 3.24.4 → 3.24.6

package/README.md

@@ -4,7 +4,7 @@
 
 ### Cut your LLM bill by <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->%. One line of TypeScript.
 
-The smart-routing SDK for <!-- br:models.chatVisible -->88<!-- /br:models.chatVisible --> models — every request goes to the cheapest model that can handle it,
+The smart-routing SDK for <!-- br:models.chatVisible -->90<!-- /br:models.chatVisible --> models — every request goes to the cheapest model that can handle it,
 paid with an API key or per-request USDC on Solana, Base or Arc. No vendor lock-in.
 
 [![npm](https://img.shields.io/npm/v/@blockrun/llm.svg?style=flat-square)](https://www.npmjs.com/package/@blockrun/llm)
@@ -56,7 +56,7 @@ console.log(r.response);         // the proof
 |                    | OpenAI SDK     | OpenRouter        | LiteLLM          | **@blockrun/llm**                                                       |
 | ------------------ | -------------- | ----------------- | ---------------- | ----------------------------------------------------------------------- |
 | **Cost routing**   | ✗ one vendor   | Manual selection  | Manual selection | **Automatic — <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% cheaper** |
-| **Models**         | GPT only       | 200+              | 100+ (BYO keys)  | **<!-- br:models.chatVisible -->88<!-- /br:models.chatVisible -->, one credential** |
+| **Models**         | GPT only       | 200+              | 100+ (BYO keys)  | **<!-- br:models.chatVisible -->90<!-- /br:models.chatVisible -->, one credential** |
 | **Free tier**      | ✗              | Rate-limited      | ✗                | **<!-- br:models.free -->7<!-- /br:models.free --> models, no signup**  |
 | **Auth**           | API key        | Account + API key | Your API keys    | **API key *or* wallet signature**                                       |
 | **Payment**        | Card + invoice | Credit card       | BYO keys         | **Account credit or USDC per-request**                                  |
@@ -1811,7 +1811,7 @@ to pick up security fixes.
 ## Frequently Asked Questions
 
 ### What is @blockrun/llm?
-A TypeScript SDK that cuts LLM costs by <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% on a realistic workload with built-in smart routing: every request goes to the cheapest of <!-- br:models.chatVisible -->88<!-- /br:models.chatVisible --> models (OpenAI, Anthropic, Google, xAI, DeepSeek, Moonshot, and more) that can handle it, paid with an account API key or per-request USDC via x402 on Solana, Base or Arc.
+A TypeScript SDK that cuts LLM costs by <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% on a realistic workload with built-in smart routing: every request goes to the cheapest of <!-- br:models.chatVisible -->90<!-- /br:models.chatVisible --> models (OpenAI, Anthropic, Google, xAI, DeepSeek, Moonshot, and more) that can handle it, paid with an account API key or per-request USDC via x402 on Solana, Base or Arc.
 
 ### What is smart routing?
 Router Core V3, bundled into the SDK: the same deterministic engine that powers ClawRouter. It analyzes each request across <!-- br:clawrouter.dimensions -->15<!-- /br:clawrouter.dimensions --> dimensions locally, in under 1ms, and picks the cheapest capable model. Use `smartChat()`, `smartChatCompletion()`, or the `blockrun/auto` model alias. See [Smart Routing](#smart-routing-router-core-v3).

package/dist/index.cjs

@@ -571,7 +571,7 @@ function withCost(body, response, bookedUsd, scheme) {
 // src/client.ts
 var import_accounts2 = require("viem/accounts");
 
-// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+8773a9954b091_rjdkm4uoakz3dfyojiwn3sv45i/node_modules/@blockrun/router-core/dist/index.js
+// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+fd6d03d035ff6_meidvh7c7i3y7wjjl6r2ymc4py/node_modules/@blockrun/router-core/dist/index.js
 var UNSEGMENTED_SCRIPT = /[؀-ۿ　-ヿ㐀-䶿一-鿿가-힯豈-﫿＀-￯]/;
 var WORD_CHAR = /[\p{L}\p{N}]/u;
 var ASCII_PHRASE = /^[a-z0-9' -]+$/;
@@ -2903,10 +2903,10 @@ function isEligible(modelId, features, maxOutputTokens, options) {
   return model.contextWindow >= (features.estimatedInputTokens + maxOutputTokens) * 1.1;
 }
 function estimatedCost(modelId, options, inputTokens, outputTokens) {
-  const price = options.modelPricing.get(modelId);
-  if (!price) return Number.POSITIVE_INFINITY;
-  if (price.flatPrice !== void 0) return price.flatPrice;
-  return (inputTokens * price.inputPrice + outputTokens * price.outputPrice) / 1e6;
+  const price2 = options.modelPricing.get(modelId);
+  if (!price2) return Number.POSITIVE_INFINITY;
+  if (price2.flatPrice !== void 0) return price2.flatPrice;
+  return (inputTokens * price2.inputPrice + outputTokens * price2.outputPrice) / 1e6;
 }
 function profileScore(modelId, options, now) {
   const profile = options.modelPerformance?.[modelId] ?? LIVE_MODEL_PROFILES[modelId] ?? HISTORICAL_MODEL_PROFILES[modelId];
@@ -5089,7 +5089,7 @@ function getCostSummary() {
 }
 
 // src/version.ts
-var SDK_VERSION = "3.24.4";
+var SDK_VERSION = "3.24.6";
 var USER_AGENT = `blockrun-ts/${SDK_VERSION}`;
 
 // src/client.ts
@@ -5157,8 +5157,15 @@ var LLMClient = class _LLMClient {
   modelPricingCache = null;
   modelPricingPromise = null;
   // Pre-auth cache: avoids the 402 round-trip on repeat requests to the same model.
-  // Key = "endpoint:model", value = cached payment header + timestamp.
+  // Key = "endpoint:model", value = cached payment header + timestamp + the
+  // length of the request body it was quoted for.
   // TTL: 1 hour — pre-auth quotes are stable server-side on that horizon.
+  //
+  // A quote prices ONE request (input size + max_tokens), and exact settles
+  // the whole signed amount, so reusing it for a smaller request overpaid and
+  // for a larger one underpaid. The gateway refuses both with a 402 (an
+  // over-signed payment since blockrun#874), so a reuse is only attempted for
+  // a body the same size as the one that was quoted.
   preAuthCache = /* @__PURE__ */ new Map();
   static PRE_AUTH_TTL_MS = 36e5;
   /**
@@ -5765,9 +5772,10 @@ var LLMClient = class _LLMClient {
     if (options?.responseFormat !== void 0) body.response_format = options.responseFormat;
     if (options?.stop !== void 0) body.stop = options.stop;
     const cacheKey2 = `/v1/chat/completions:${requestModel}`;
+    const requestBody = JSON.stringify(body);
     const cached = this.preAuthCache.get(cacheKey2);
     const now = Date.now();
-    if (cached && now - cached.cachedAt < _LLMClient.PRE_AUTH_TTL_MS) {
+    if (cached && now - cached.cachedAt < _LLMClient.PRE_AUTH_TTL_MS && cached.bodyLength === requestBody.length) {
       let signed;
       try {
         signed = await this.signPayment(cached.paymentHeader);
@@ -5836,7 +5844,7 @@ var LLMClient = class _LLMClient {
       }
     }
     if (!paymentHeader) throw new PaymentError("402 response but no payment requirements found");
-    this.preAuthCache.set(cacheKey2, { paymentHeader, cachedAt: now });
+    this.preAuthCache.set(cacheKey2, { paymentHeader, cachedAt: now, bodyLength: requestBody.length });
     const { paymentPayload, costUsd } = await this.signPayment(paymentHeader);
     const streamResp = await this.fetchWithTimeout(url, {
       method: "POST",

package/dist/index.js

@@ -440,7 +440,7 @@ function withCost(body, response, bookedUsd, scheme) {
 // src/client.ts
 import { privateKeyToAccount } from "viem/accounts";
 
-// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+8773a9954b091_rjdkm4uoakz3dfyojiwn3sv45i/node_modules/@blockrun/router-core/dist/index.js
+// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+fd6d03d035ff6_meidvh7c7i3y7wjjl6r2ymc4py/node_modules/@blockrun/router-core/dist/index.js
 var UNSEGMENTED_SCRIPT = /[؀-ۿ　-ヿ㐀-䶿一-鿿가-힯豈-﫿＀-￯]/;
 var WORD_CHAR = /[\p{L}\p{N}]/u;
 var ASCII_PHRASE = /^[a-z0-9' -]+$/;
@@ -2772,10 +2772,10 @@ function isEligible(modelId, features, maxOutputTokens, options) {
   return model.contextWindow >= (features.estimatedInputTokens + maxOutputTokens) * 1.1;
 }
 function estimatedCost(modelId, options, inputTokens, outputTokens) {
-  const price = options.modelPricing.get(modelId);
-  if (!price) return Number.POSITIVE_INFINITY;
-  if (price.flatPrice !== void 0) return price.flatPrice;
-  return (inputTokens * price.inputPrice + outputTokens * price.outputPrice) / 1e6;
+  const price2 = options.modelPricing.get(modelId);
+  if (!price2) return Number.POSITIVE_INFINITY;
+  if (price2.flatPrice !== void 0) return price2.flatPrice;
+  return (inputTokens * price2.inputPrice + outputTokens * price2.outputPrice) / 1e6;
 }
 function profileScore(modelId, options, now) {
   const profile = options.modelPerformance?.[modelId] ?? LIVE_MODEL_PROFILES[modelId] ?? HISTORICAL_MODEL_PROFILES[modelId];
@@ -4958,7 +4958,7 @@ function getCostSummary() {
 }
 
 // src/version.ts
-var SDK_VERSION = "3.24.4";
+var SDK_VERSION = "3.24.6";
 var USER_AGENT = `blockrun-ts/${SDK_VERSION}`;
 
 // src/client.ts
@@ -5026,8 +5026,15 @@ var LLMClient = class _LLMClient {
   modelPricingCache = null;
   modelPricingPromise = null;
   // Pre-auth cache: avoids the 402 round-trip on repeat requests to the same model.
-  // Key = "endpoint:model", value = cached payment header + timestamp.
+  // Key = "endpoint:model", value = cached payment header + timestamp + the
+  // length of the request body it was quoted for.
   // TTL: 1 hour — pre-auth quotes are stable server-side on that horizon.
+  //
+  // A quote prices ONE request (input size + max_tokens), and exact settles
+  // the whole signed amount, so reusing it for a smaller request overpaid and
+  // for a larger one underpaid. The gateway refuses both with a 402 (an
+  // over-signed payment since blockrun#874), so a reuse is only attempted for
+  // a body the same size as the one that was quoted.
   preAuthCache = /* @__PURE__ */ new Map();
   static PRE_AUTH_TTL_MS = 36e5;
   /**
@@ -5634,9 +5641,10 @@ var LLMClient = class _LLMClient {
     if (options?.responseFormat !== void 0) body.response_format = options.responseFormat;
     if (options?.stop !== void 0) body.stop = options.stop;
     const cacheKey2 = `/v1/chat/completions:${requestModel}`;
+    const requestBody = JSON.stringify(body);
     const cached = this.preAuthCache.get(cacheKey2);
     const now = Date.now();
-    if (cached && now - cached.cachedAt < _LLMClient.PRE_AUTH_TTL_MS) {
+    if (cached && now - cached.cachedAt < _LLMClient.PRE_AUTH_TTL_MS && cached.bodyLength === requestBody.length) {
       let signed;
       try {
         signed = await this.signPayment(cached.paymentHeader);
@@ -5705,7 +5713,7 @@ var LLMClient = class _LLMClient {
       }
     }
     if (!paymentHeader) throw new PaymentError("402 response but no payment requirements found");
-    this.preAuthCache.set(cacheKey2, { paymentHeader, cachedAt: now });
+    this.preAuthCache.set(cacheKey2, { paymentHeader, cachedAt: now, bodyLength: requestBody.length });
     const { paymentPayload, costUsd } = await this.signPayment(paymentHeader);
     const streamResp = await this.fetchWithTimeout(url, {
       method: "POST",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@blockrun/llm",
-  "version": "3.24.4",
+  "version": "3.24.6",
   "type": "module",
   "description": "TypeScript SDK for BlockRun - every frontier model behind one OpenAI-compatible client. Pay with a BlockRun API key or per-call in USDC. Smart routing picks the cheapest capable model. No rate limits.",
   "main": "dist/index.cjs",
@@ -58,7 +58,7 @@
   },
   "devDependencies": {
     "@blockrun/core": "^0.1.0",
-    "@blockrun/router-core": "https://codeload.github.com/BlockRunAI/router-core/tar.gz/8773a9954b091bd0090acdfec67c75a006c0e156",
+    "@blockrun/router-core": "https://codeload.github.com/BlockRunAI/router-core/tar.gz/fd6d03d035ff676642ba1683c333d929939f8a00",
     "@eslint/js": "^10.0.1",
     "@solana/kit": "^5.5.1",
     "@types/node": "^26.4.1",