@llmgateway/ai-sdk-provider 1.0.3 → 2.0.0

package/README.md

@@ -32,7 +32,7 @@ import { llmgateway, createLLMGateway } from '@llmgateway/ai-sdk-provider';
 import { generateText } from 'ai';
 
 const llmgateway = createLLMGateway({
-  apiKey: process.env.LLMGATEWAY_API_KEY,
+  apiKey: process.env.LLM_GATEWAY_API_KEY,
 });
 
 const { text } = await generateText({

package/dist/index.d.mts

@@ -1,15 +1,55 @@
-import { LanguageModelV1 } from '@ai-sdk/provider';
-export { LanguageModelV1 } from '@ai-sdk/provider';
+import { LanguageModelV2, LanguageModelV2CallOptions, LanguageModelV2Content, LanguageModelV2FinishReason, LanguageModelV2Usage, LanguageModelV2CallWarning, LanguageModelV2ResponseMetadata, SharedV2Headers, LanguageModelV2StreamPart } from '@ai-sdk/provider';
+export { LanguageModelV2, LanguageModelV2Prompt } from '@ai-sdk/provider';
+
+type LLMGatewayChatModelId = string;
+type LLMGatewayChatSettings = {
+    /**
+  Modify the likelihood of specified tokens appearing in the completion.
+  
+  Accepts a JSON object that maps tokens (specified by their token ID in
+  the GPT tokenizer) to an associated bias value from -100 to 100. You
+  can use this tokenizer tool to convert text to token IDs. Mathematically,
+  the bias is added to the logits generated by the model prior to sampling.
+  The exact effect will vary per model, but values between -1 and 1 should
+  decrease or increase likelihood of selection; values like -100 or 100
+  should result in a ban or exclusive selection of the relevant token.
+  
+  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
+  token from being generated.
+  */
+    logitBias?: Record<number, number>;
+    /**
+  Return the log probabilities of the tokens. Including logprobs will increase
+  the response size and can slow down response times. However, it can
+  be useful to understand better how the model is behaving.
+  
+  Setting to true will return the log probabilities of the tokens that
+  were generated.
+  
+  Setting to a number will return the log probabilities of the top n
+  tokens that were generated.
+  */
+    logprobs?: boolean | number;
+    /**
+  Whether to enable parallel function calling during tool use. Default to true.
+     */
+    parallelToolCalls?: boolean;
+    /**
+  A unique identifier representing your end-user, which can help LLMGateway to
+  monitor and detect abuse. Learn more.
+  */
+    user?: string;
+} & LLMGatewaySharedSettings;
 
-type LLMGatewayLanguageModel = LanguageModelV1;
 type LLMGatewayProviderOptions = {
     models?: string[];
     /**
-     * https://llmgateway.io/docs/use-cases/reasoning-tokens
+     * Reasoning configuration for supported models.
      * One of `max_tokens` or `effort` is required.
      * If `exclude` is true, reasoning will be removed from the response. Default is false.
      */
     reasoning?: {
+        enabled?: boolean;
         exclude?: boolean;
     } & ({
         max_tokens: number;
@@ -30,7 +70,6 @@ type LLMGatewaySharedSettings = LLMGatewayProviderOptions & {
     extraBody?: Record<string, unknown>;
     /**
      * Enable usage accounting to get detailed token usage information.
-     * https://llmgateway.io/docs/use-cases/usage-accounting
      */
     usage?: {
         /**
@@ -41,7 +80,6 @@ type LLMGatewaySharedSettings = LLMGatewayProviderOptions & {
 };
 /**
  * Usage accounting response
- * @see https://llmgateway.io/docs/use-cases/usage-accounting
  */
 type LLMGatewayUsageAccounting = {
     promptTokens: number;
@@ -54,6 +92,9 @@ type LLMGatewayUsageAccounting = {
     };
     totalTokens: number;
     cost?: number;
+    costDetails: {
+        upstreamInferenceCost: number;
+    };
 };
 
 type LLMGatewayCompletionModelId = string;
@@ -91,46 +132,6 @@ type LLMGatewayCompletionSettings = {
     suffix?: string;
 } & LLMGatewaySharedSettings;
 
-type LLMGatewayChatModelId = string;
-type LLMGatewayChatSettings = {
-    /**
-  Modify the likelihood of specified tokens appearing in the completion.
-  
-  Accepts a JSON object that maps tokens (specified by their token ID in
-  the GPT tokenizer) to an associated bias value from -100 to 100. You
-  can use this tokenizer tool to convert text to token IDs. Mathematically,
-  the bias is added to the logits generated by the model prior to sampling.
-  The exact effect will vary per model, but values between -1 and 1 should
-  decrease or increase likelihood of selection; values like -100 or 100
-  should result in a ban or exclusive selection of the relevant token.
-  
-  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
-  token from being generated.
-  */
-    logitBias?: Record<number, number>;
-    /**
-  Return the log probabilities of the tokens. Including logprobs will increase
-  the response size and can slow down response times. However, it can
-  be useful to better understand how the model is behaving.
-  
-  Setting to true will return the log probabilities of the tokens that
-  were generated.
-  
-  Setting to a number will return the log probabilities of the top n
-  tokens that were generated.
-  */
-    logprobs?: boolean | number;
-    /**
-  Whether to enable parallel function calling during tool use. Default to true.
-     */
-    parallelToolCalls?: boolean;
-    /**
-  A unique identifier representing your end-user, which can help LLMGateway to
-  monitor and detect abuse. Learn more.
-  */
-    user?: string;
-} & LLMGatewaySharedSettings;
-
 type LLMGatewayChatConfig = {
     provider: string;
     compatibility: 'strict' | 'compatible';
@@ -142,19 +143,45 @@ type LLMGatewayChatConfig = {
     fetch?: typeof fetch;
     extraBody?: Record<string, unknown>;
 };
-type DoGenerateOutput = Awaited<ReturnType<LanguageModelV1['doGenerate']>>;
-type DoStreamOutput = Awaited<ReturnType<LanguageModelV1['doStream']>>;
-declare class LLMGatewayChatLanguageModel implements LanguageModelV1 {
-    readonly specificationVersion = "v1";
-    readonly defaultObjectGenerationMode = "tool";
+declare class LLMGatewayChatLanguageModel implements LanguageModelV2 {
+    readonly specificationVersion: "v2";
+    readonly provider = "llmgateway";
+    readonly defaultObjectGenerationMode: "tool";
     readonly modelId: LLMGatewayChatModelId;
+    readonly supportedUrls: Record<string, RegExp[]>;
     readonly settings: LLMGatewayChatSettings;
     private readonly config;
     constructor(modelId: LLMGatewayChatModelId, settings: LLMGatewayChatSettings, config: LLMGatewayChatConfig);
-    get provider(): string;
     private getArgs;
-    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<DoGenerateOutput>;
-    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<DoStreamOutput>;
+    doGenerate(options: LanguageModelV2CallOptions): Promise<{
+        content: Array<LanguageModelV2Content>;
+        finishReason: LanguageModelV2FinishReason;
+        usage: LanguageModelV2Usage;
+        warnings: Array<LanguageModelV2CallWarning>;
+        providerMetadata?: {
+            llmgateway: {
+                usage: LLMGatewayUsageAccounting;
+            };
+        };
+        request?: {
+            body?: unknown;
+        };
+        response?: LanguageModelV2ResponseMetadata & {
+            headers?: SharedV2Headers;
+            body?: unknown;
+        };
+    }>;
+    doStream(options: LanguageModelV2CallOptions): Promise<{
+        stream: ReadableStream<LanguageModelV2StreamPart>;
+        warnings: Array<LanguageModelV2CallWarning>;
+        request?: {
+            body?: unknown;
+        };
+        response?: LanguageModelV2ResponseMetadata & {
+            headers?: SharedV2Headers;
+            body?: unknown;
+        };
+    }>;
 }
 
 type LLMGatewayCompletionConfig = {
@@ -168,20 +195,21 @@ type LLMGatewayCompletionConfig = {
     fetch?: typeof fetch;
     extraBody?: Record<string, unknown>;
 };
-declare class LLMGatewayCompletionLanguageModel implements LanguageModelV1 {
-    readonly specificationVersion = "v1";
-    readonly defaultObjectGenerationMode: undefined;
+declare class LLMGatewayCompletionLanguageModel implements LanguageModelV2 {
+    readonly specificationVersion: "v2";
+    readonly provider = "llmgateway";
     readonly modelId: LLMGatewayCompletionModelId;
+    readonly supportedUrls: Record<string, RegExp[]>;
+    readonly defaultObjectGenerationMode: undefined;
     readonly settings: LLMGatewayCompletionSettings;
     private readonly config;
     constructor(modelId: LLMGatewayCompletionModelId, settings: LLMGatewayCompletionSettings, config: LLMGatewayCompletionConfig);
-    get provider(): string;
     private getArgs;
-    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
-    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+    doGenerate(options: LanguageModelV2CallOptions): Promise<Awaited<ReturnType<LanguageModelV2['doGenerate']>>>;
+    doStream(options: LanguageModelV2CallOptions): Promise<Awaited<ReturnType<LanguageModelV2['doStream']>>>;
 }
 
-interface LLMGatewayProvider {
+interface LLMGatewayProvider extends LanguageModelV2 {
     (modelId: LLMGatewayChatModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
     (modelId: LLMGatewayChatModelId, settings?: LLMGatewayChatSettings): LLMGatewayChatLanguageModel;
     languageModel(modelId: LLMGatewayChatModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
@@ -201,6 +229,10 @@ interface LLMGatewayProviderSettings {
        */
     baseURL?: string;
     /**
+  @deprecated Use `baseURL` instead.
+       */
+    baseUrl?: string;
+    /**
   API key for authenticating requests.
        */
     apiKey?: string;
@@ -243,8 +275,8 @@ declare class LLMGateway {
      */
     readonly baseURL: string;
     /**
-  API key that is being send using the `Authorization` header.
-  It defaults to the `LLMGATEWAY_API_KEY` environment variable.
+  API key that is being sent using the `Authorization` header.
+  It defaults to the `LLM_GATEWAY_API_KEY` environment variable.
    */
     readonly apiKey?: string;
     /**
@@ -260,4 +292,4 @@ declare class LLMGateway {
     completion(modelId: LLMGatewayCompletionModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
 }
 
-export { LLMGateway, type LLMGatewayCompletionSettings, type LLMGatewayLanguageModel, type LLMGatewayProvider, type LLMGatewayProviderOptions, type LLMGatewayProviderSettings, type LLMGatewaySharedSettings, type LLMGatewayUsageAccounting, createLLMGateway, llmgateway };
+export { LLMGateway, type LLMGatewayCompletionSettings, type LLMGatewayProvider, type LLMGatewayProviderOptions, type LLMGatewayProviderSettings, type LLMGatewaySharedSettings, type LLMGatewayUsageAccounting, createLLMGateway, llmgateway };

package/dist/index.d.ts

@@ -1,15 +1,55 @@
-import { LanguageModelV1 } from '@ai-sdk/provider';
-export { LanguageModelV1 } from '@ai-sdk/provider';
+import { LanguageModelV2, LanguageModelV2CallOptions, LanguageModelV2Content, LanguageModelV2FinishReason, LanguageModelV2Usage, LanguageModelV2CallWarning, LanguageModelV2ResponseMetadata, SharedV2Headers, LanguageModelV2StreamPart } from '@ai-sdk/provider';
+export { LanguageModelV2, LanguageModelV2Prompt } from '@ai-sdk/provider';
+
+type LLMGatewayChatModelId = string;
+type LLMGatewayChatSettings = {
+    /**
+  Modify the likelihood of specified tokens appearing in the completion.
+  
+  Accepts a JSON object that maps tokens (specified by their token ID in
+  the GPT tokenizer) to an associated bias value from -100 to 100. You
+  can use this tokenizer tool to convert text to token IDs. Mathematically,
+  the bias is added to the logits generated by the model prior to sampling.
+  The exact effect will vary per model, but values between -1 and 1 should
+  decrease or increase likelihood of selection; values like -100 or 100
+  should result in a ban or exclusive selection of the relevant token.
+  
+  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
+  token from being generated.
+  */
+    logitBias?: Record<number, number>;
+    /**
+  Return the log probabilities of the tokens. Including logprobs will increase
+  the response size and can slow down response times. However, it can
+  be useful to understand better how the model is behaving.
+  
+  Setting to true will return the log probabilities of the tokens that
+  were generated.
+  
+  Setting to a number will return the log probabilities of the top n
+  tokens that were generated.
+  */
+    logprobs?: boolean | number;
+    /**
+  Whether to enable parallel function calling during tool use. Default to true.
+     */
+    parallelToolCalls?: boolean;
+    /**
+  A unique identifier representing your end-user, which can help LLMGateway to
+  monitor and detect abuse. Learn more.
+  */
+    user?: string;
+} & LLMGatewaySharedSettings;
 
-type LLMGatewayLanguageModel = LanguageModelV1;
 type LLMGatewayProviderOptions = {
     models?: string[];
     /**
-     * https://llmgateway.io/docs/use-cases/reasoning-tokens
+     * Reasoning configuration for supported models.
      * One of `max_tokens` or `effort` is required.
      * If `exclude` is true, reasoning will be removed from the response. Default is false.
      */
     reasoning?: {
+        enabled?: boolean;
         exclude?: boolean;
     } & ({
         max_tokens: number;
@@ -30,7 +70,6 @@ type LLMGatewaySharedSettings = LLMGatewayProviderOptions & {
     extraBody?: Record<string, unknown>;
     /**
      * Enable usage accounting to get detailed token usage information.
-     * https://llmgateway.io/docs/use-cases/usage-accounting
      */
     usage?: {
         /**
@@ -41,7 +80,6 @@ type LLMGatewaySharedSettings = LLMGatewayProviderOptions & {
 };
 /**
  * Usage accounting response
- * @see https://llmgateway.io/docs/use-cases/usage-accounting
  */
 type LLMGatewayUsageAccounting = {
     promptTokens: number;
@@ -54,6 +92,9 @@ type LLMGatewayUsageAccounting = {
     };
     totalTokens: number;
     cost?: number;
+    costDetails: {
+        upstreamInferenceCost: number;
+    };
 };
 
 type LLMGatewayCompletionModelId = string;
@@ -91,46 +132,6 @@ type LLMGatewayCompletionSettings = {
     suffix?: string;
 } & LLMGatewaySharedSettings;
 
-type LLMGatewayChatModelId = string;
-type LLMGatewayChatSettings = {
-    /**
-  Modify the likelihood of specified tokens appearing in the completion.
-  
-  Accepts a JSON object that maps tokens (specified by their token ID in
-  the GPT tokenizer) to an associated bias value from -100 to 100. You
-  can use this tokenizer tool to convert text to token IDs. Mathematically,
-  the bias is added to the logits generated by the model prior to sampling.
-  The exact effect will vary per model, but values between -1 and 1 should
-  decrease or increase likelihood of selection; values like -100 or 100
-  should result in a ban or exclusive selection of the relevant token.
-  
-  As an example, you can pass {"50256": -100} to prevent the <|endoftext|>
-  token from being generated.
-  */
-    logitBias?: Record<number, number>;
-    /**
-  Return the log probabilities of the tokens. Including logprobs will increase
-  the response size and can slow down response times. However, it can
-  be useful to better understand how the model is behaving.
-  
-  Setting to true will return the log probabilities of the tokens that
-  were generated.
-  
-  Setting to a number will return the log probabilities of the top n
-  tokens that were generated.
-  */
-    logprobs?: boolean | number;
-    /**
-  Whether to enable parallel function calling during tool use. Default to true.
-     */
-    parallelToolCalls?: boolean;
-    /**
-  A unique identifier representing your end-user, which can help LLMGateway to
-  monitor and detect abuse. Learn more.
-  */
-    user?: string;
-} & LLMGatewaySharedSettings;
-
 type LLMGatewayChatConfig = {
     provider: string;
     compatibility: 'strict' | 'compatible';
@@ -142,19 +143,45 @@ type LLMGatewayChatConfig = {
     fetch?: typeof fetch;
     extraBody?: Record<string, unknown>;
 };
-type DoGenerateOutput = Awaited<ReturnType<LanguageModelV1['doGenerate']>>;
-type DoStreamOutput = Awaited<ReturnType<LanguageModelV1['doStream']>>;
-declare class LLMGatewayChatLanguageModel implements LanguageModelV1 {
-    readonly specificationVersion = "v1";
-    readonly defaultObjectGenerationMode = "tool";
+declare class LLMGatewayChatLanguageModel implements LanguageModelV2 {
+    readonly specificationVersion: "v2";
+    readonly provider = "llmgateway";
+    readonly defaultObjectGenerationMode: "tool";
     readonly modelId: LLMGatewayChatModelId;
+    readonly supportedUrls: Record<string, RegExp[]>;
     readonly settings: LLMGatewayChatSettings;
     private readonly config;
     constructor(modelId: LLMGatewayChatModelId, settings: LLMGatewayChatSettings, config: LLMGatewayChatConfig);
-    get provider(): string;
     private getArgs;
-    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<DoGenerateOutput>;
-    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<DoStreamOutput>;
+    doGenerate(options: LanguageModelV2CallOptions): Promise<{
+        content: Array<LanguageModelV2Content>;
+        finishReason: LanguageModelV2FinishReason;
+        usage: LanguageModelV2Usage;
+        warnings: Array<LanguageModelV2CallWarning>;
+        providerMetadata?: {
+            llmgateway: {
+                usage: LLMGatewayUsageAccounting;
+            };
+        };
+        request?: {
+            body?: unknown;
+        };
+        response?: LanguageModelV2ResponseMetadata & {
+            headers?: SharedV2Headers;
+            body?: unknown;
+        };
+    }>;
+    doStream(options: LanguageModelV2CallOptions): Promise<{
+        stream: ReadableStream<LanguageModelV2StreamPart>;
+        warnings: Array<LanguageModelV2CallWarning>;
+        request?: {
+            body?: unknown;
+        };
+        response?: LanguageModelV2ResponseMetadata & {
+            headers?: SharedV2Headers;
+            body?: unknown;
+        };
+    }>;
 }
 
 type LLMGatewayCompletionConfig = {
@@ -168,20 +195,21 @@ type LLMGatewayCompletionConfig = {
     fetch?: typeof fetch;
     extraBody?: Record<string, unknown>;
 };
-declare class LLMGatewayCompletionLanguageModel implements LanguageModelV1 {
-    readonly specificationVersion = "v1";
-    readonly defaultObjectGenerationMode: undefined;
+declare class LLMGatewayCompletionLanguageModel implements LanguageModelV2 {
+    readonly specificationVersion: "v2";
+    readonly provider = "llmgateway";
     readonly modelId: LLMGatewayCompletionModelId;
+    readonly supportedUrls: Record<string, RegExp[]>;
+    readonly defaultObjectGenerationMode: undefined;
     readonly settings: LLMGatewayCompletionSettings;
     private readonly config;
     constructor(modelId: LLMGatewayCompletionModelId, settings: LLMGatewayCompletionSettings, config: LLMGatewayCompletionConfig);
-    get provider(): string;
     private getArgs;
-    doGenerate(options: Parameters<LanguageModelV1['doGenerate']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doGenerate']>>>;
-    doStream(options: Parameters<LanguageModelV1['doStream']>[0]): Promise<Awaited<ReturnType<LanguageModelV1['doStream']>>>;
+    doGenerate(options: LanguageModelV2CallOptions): Promise<Awaited<ReturnType<LanguageModelV2['doGenerate']>>>;
+    doStream(options: LanguageModelV2CallOptions): Promise<Awaited<ReturnType<LanguageModelV2['doStream']>>>;
 }
 
-interface LLMGatewayProvider {
+interface LLMGatewayProvider extends LanguageModelV2 {
     (modelId: LLMGatewayChatModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
     (modelId: LLMGatewayChatModelId, settings?: LLMGatewayChatSettings): LLMGatewayChatLanguageModel;
     languageModel(modelId: LLMGatewayChatModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
@@ -201,6 +229,10 @@ interface LLMGatewayProviderSettings {
        */
     baseURL?: string;
     /**
+  @deprecated Use `baseURL` instead.
+       */
+    baseUrl?: string;
+    /**
   API key for authenticating requests.
        */
     apiKey?: string;
@@ -243,8 +275,8 @@ declare class LLMGateway {
      */
     readonly baseURL: string;
     /**
-  API key that is being send using the `Authorization` header.
-  It defaults to the `LLMGATEWAY_API_KEY` environment variable.
+  API key that is being sent using the `Authorization` header.
+  It defaults to the `LLM_GATEWAY_API_KEY` environment variable.
    */
     readonly apiKey?: string;
     /**
@@ -260,4 +292,4 @@ declare class LLMGateway {
     completion(modelId: LLMGatewayCompletionModelId, settings?: LLMGatewayCompletionSettings): LLMGatewayCompletionLanguageModel;
 }
 
-export { LLMGateway, type LLMGatewayCompletionSettings, type LLMGatewayLanguageModel, type LLMGatewayProvider, type LLMGatewayProviderOptions, type LLMGatewayProviderSettings, type LLMGatewaySharedSettings, type LLMGatewayUsageAccounting, createLLMGateway, llmgateway };
+export { LLMGateway, type LLMGatewayCompletionSettings, type LLMGatewayProvider, type LLMGatewayProviderOptions, type LLMGatewayProviderSettings, type LLMGatewaySharedSettings, type LLMGatewayUsageAccounting, createLLMGateway, llmgateway };