@cencori/ai-sdk 0.2.1

package/README.md

@@ -0,0 +1,106 @@
+# @cencori/ai-sdk
+
+The Cencori AI SDK — the infrastructure layer for AI applications. Works with [Vercel AI SDK](https://github.com/vercel/ai), TanStack AI, and more.
+
+## Installation
+
+```bash
+npm install @cencori/ai-sdk ai
+```
+
+## Quick Start
+
+```typescript
+import { cencori } from '@cencori/ai-sdk';
+import { streamText } from 'ai';
+
+const result = await streamText({
+  model: cencori('gemini-2.5-flash'),
+  messages: [{ role: 'user', content: 'Hello!' }]
+});
+
+for await (const chunk of result.textStream) {
+  process.stdout.write(chunk);
+}
+```
+
+## Usage with Next.js App Router
+
+```typescript
+// app/api/chat/route.ts
+import { cencori } from '@cencori/ai-sdk';
+import { streamText } from 'ai';
+
+export async function POST(req: Request) {
+  const { messages } = await req.json();
+
+  const result = await streamText({
+    model: cencori('gemini-2.5-flash'),
+    messages
+  });
+
+  return result.toUIMessageStreamResponse();
+}
+```
+
+## Configuration
+
+### Environment Variable
+
+Set the `CENCORI_API_KEY` environment variable:
+
+```bash
+CENCORI_API_KEY=csk_your_key_here
+```
+
+### Custom Configuration
+
+```typescript
+import { createCencori } from '@cencori/ai-sdk';
+
+const cencori = createCencori({
+  apiKey: 'csk_your_key_here',
+  baseUrl: 'https://cencori.com', // optional
+});
+
+const result = await streamText({
+  model: cencori('gpt-4o'),
+  messages: [{ role: 'user', content: 'Hello!' }]
+});
+```
+
+## Supported Models
+
+Use any model supported by Cencori:
+
+| Provider | Models |
+|----------|--------|
+| OpenAI | `gpt-4o`, `gpt-4o-mini`, `o1` |
+| Anthropic | `claude-3-opus`, `claude-3-5-sonnet`, `claude-3-haiku` |
+| Google | `gemini-2.5-flash`, `gemini-2.0-flash`, `gemini-3-pro` |
+| xAI | `grok-4`, `grok-3` |
+| Mistral | `mistral-large`, `codestral` |
+| DeepSeek | `deepseek-v3.2`, `deepseek-reasoner` |
+| + More | Groq, Cohere, Perplexity, Together, Meta, Qwen, HuggingFace |
+
+## Why Cencori?
+
+Unlike raw AI SDKs, Cencori gives you:
+
+- 🔒 **Security** — PII filtering, jailbreak detection, content moderation
+- 📊 **Observability** — Request logs, latency metrics, cost tracking
+- 💰 **Cost Control** — Budgets, alerts, per-route analytics
+- 🔌 **Multi-Provider** — One API key for OpenAI, Claude, Gemini, and more
+
+## Features
+
+- ✅ Drop-in Vercel AI SDK compatibility
+- ✅ Works with `streamText()`, `generateText()`, `useChat()`
+- ✅ Built-in content safety filtering
+- ✅ Rate limiting protection
+- ✅ Full analytics in Cencori dashboard
+- ✅ Multi-provider support with one API key
+
+## License
+
+MIT © FohnAI

package/dist/index.d.mts

@@ -0,0 +1,118 @@
+import { LanguageModelV3, LanguageModelV3CallOptions, LanguageModelV3GenerateResult, LanguageModelV3StreamResult } from '@ai-sdk/provider';
+
+/**
+ * Cencori Chat Language Model
+ *
+ * Implements the Vercel AI SDK's LanguageModelV3 interface (AI SDK v6 compatible)
+ */
+
+interface CencoriChatModelSettings {
+    apiKey: string;
+    baseUrl: string;
+    headers?: Record<string, string>;
+    userId?: string;
+}
+declare class CencoriChatLanguageModel implements LanguageModelV3 {
+    readonly specificationVersion: "v3";
+    readonly provider = "cencori";
+    readonly modelId: string;
+    readonly supportedUrls: Record<string, RegExp[]>;
+    private readonly settings;
+    constructor(modelId: string, settings: CencoriChatModelSettings);
+    private getHeaders;
+    private convertMessages;
+    private mapFinishReason;
+    private buildUsage;
+    doGenerate(options: LanguageModelV3CallOptions): Promise<LanguageModelV3GenerateResult>;
+    doStream(options: LanguageModelV3CallOptions): Promise<LanguageModelV3StreamResult>;
+}
+
+/**
+ * Types for Cencori AI Provider
+ */
+interface CencoriProviderSettings {
+    /**
+     * Cencori API key (csk_ or cpk_ prefix)
+     */
+    apiKey?: string;
+    /**
+     * Base URL for the Cencori API
+     * @default 'https://cencori.com'
+     */
+    baseUrl?: string;
+    /**
+     * Custom headers to include in requests
+     */
+    headers?: Record<string, string>;
+}
+interface CencoriChatSettings {
+    /**
+     * Optional user ID for rate limiting and analytics
+     */
+    userId?: string;
+}
+
+/**
+ * Cencori AI Provider for Vercel AI SDK
+ *
+ * Use Cencori with streamText(), generateText(), and useChat()
+ */
+
+interface CencoriProvider {
+    /**
+     * Create a Cencori chat model for use with Vercel AI SDK
+     *
+     * @param modelId - The model ID (e.g., 'gemini-2.5-flash', 'gpt-4o', 'claude-3-opus')
+     * @param settings - Optional model-specific settings
+     * @returns A LanguageModelV1 compatible model
+     *
+     * @example
+     * import { cencori } from '@cencori/ai-sdk';
+     * import { streamText } from 'ai';
+     *
+     * const result = await streamText({
+     *   model: cencori('gemini-2.5-flash'),
+     *   messages: [{ role: 'user', content: 'Hello!' }]
+     * });
+     */
+    (modelId: string, settings?: CencoriChatSettings): CencoriChatLanguageModel;
+    /**
+     * Create a chat model (alias for the provider function)
+     */
+    chat: (modelId: string, settings?: CencoriChatSettings) => CencoriChatLanguageModel;
+}
+/**
+ * Create a Cencori provider instance
+ *
+ * @param options - Provider configuration options
+ * @returns A Cencori provider
+ *
+ * @example
+ * import { createCencori } from '@cencori/ai-sdk';
+ *
+ * const cencori = createCencori({
+ *   apiKey: process.env.CENCORI_API_KEY
+ * });
+ *
+ * const result = await streamText({
+ *   model: cencori('gemini-2.5-flash'),
+ *   messages: [{ role: 'user', content: 'Hello!' }]
+ * });
+ */
+declare function createCencori(options?: CencoriProviderSettings): CencoriProvider;
+/**
+ * Default Cencori provider instance
+ * Uses CENCORI_API_KEY environment variable (lazy initialization)
+ *
+ * @example
+ * import { cencori } from '@cencori/ai-sdk';
+ * import { streamText } from 'ai';
+ *
+ * const result = await streamText({
+ *   model: cencori('gemini-2.5-flash'),
+ *   messages: [{ role: 'user', content: 'Hello!' }]
+ * });
+ */
+declare const cencori: CencoriProvider;
+
+export { CencoriChatLanguageModel, type CencoriChatSettings, type CencoriProvider, type CencoriProviderSettings, cencori, createCencori };

package/dist/index.d.ts

@@ -0,0 +1,118 @@
+import { LanguageModelV3, LanguageModelV3CallOptions, LanguageModelV3GenerateResult, LanguageModelV3StreamResult } from '@ai-sdk/provider';
+
+/**
+ * Cencori Chat Language Model
+ *
+ * Implements the Vercel AI SDK's LanguageModelV3 interface (AI SDK v6 compatible)
+ */
+
+interface CencoriChatModelSettings {
+    apiKey: string;
+    baseUrl: string;
+    headers?: Record<string, string>;
+    userId?: string;
+}
+declare class CencoriChatLanguageModel implements LanguageModelV3 {
+    readonly specificationVersion: "v3";
+    readonly provider = "cencori";
+    readonly modelId: string;
+    readonly supportedUrls: Record<string, RegExp[]>;
+    private readonly settings;
+    constructor(modelId: string, settings: CencoriChatModelSettings);
+    private getHeaders;
+    private convertMessages;
+    private mapFinishReason;
+    private buildUsage;
+    doGenerate(options: LanguageModelV3CallOptions): Promise<LanguageModelV3GenerateResult>;
+    doStream(options: LanguageModelV3CallOptions): Promise<LanguageModelV3StreamResult>;
+}
+
+/**
+ * Types for Cencori AI Provider
+ */
+interface CencoriProviderSettings {
+    /**
+     * Cencori API key (csk_ or cpk_ prefix)
+     */
+    apiKey?: string;
+    /**
+     * Base URL for the Cencori API
+     * @default 'https://cencori.com'
+     */
+    baseUrl?: string;
+    /**
+     * Custom headers to include in requests
+     */
+    headers?: Record<string, string>;
+}
+interface CencoriChatSettings {
+    /**
+     * Optional user ID for rate limiting and analytics
+     */
+    userId?: string;
+}
+
+/**
+ * Cencori AI Provider for Vercel AI SDK
+ *
+ * Use Cencori with streamText(), generateText(), and useChat()
+ */
+
+interface CencoriProvider {
+    /**
+     * Create a Cencori chat model for use with Vercel AI SDK
+     *
+     * @param modelId - The model ID (e.g., 'gemini-2.5-flash', 'gpt-4o', 'claude-3-opus')
+     * @param settings - Optional model-specific settings
+     * @returns A LanguageModelV1 compatible model
+     *
+     * @example
+     * import { cencori } from '@cencori/ai-sdk';
+     * import { streamText } from 'ai';
+     *
+     * const result = await streamText({
+     *   model: cencori('gemini-2.5-flash'),
+     *   messages: [{ role: 'user', content: 'Hello!' }]
+     * });
+     */
+    (modelId: string, settings?: CencoriChatSettings): CencoriChatLanguageModel;
+    /**
+     * Create a chat model (alias for the provider function)
+     */
+    chat: (modelId: string, settings?: CencoriChatSettings) => CencoriChatLanguageModel;
+}
+/**
+ * Create a Cencori provider instance
+ *
+ * @param options - Provider configuration options
+ * @returns A Cencori provider
+ *
+ * @example
+ * import { createCencori } from '@cencori/ai-sdk';
+ *
+ * const cencori = createCencori({
+ *   apiKey: process.env.CENCORI_API_KEY
+ * });
+ *
+ * const result = await streamText({
+ *   model: cencori('gemini-2.5-flash'),
+ *   messages: [{ role: 'user', content: 'Hello!' }]
+ * });
+ */
+declare function createCencori(options?: CencoriProviderSettings): CencoriProvider;
+/**
+ * Default Cencori provider instance
+ * Uses CENCORI_API_KEY environment variable (lazy initialization)
+ *
+ * @example
+ * import { cencori } from '@cencori/ai-sdk';
+ * import { streamText } from 'ai';
+ *
+ * const result = await streamText({
+ *   model: cencori('gemini-2.5-flash'),
+ *   messages: [{ role: 'user', content: 'Hello!' }]
+ * });
+ */
+declare const cencori: CencoriProvider;
+
+export { CencoriChatLanguageModel, type CencoriChatSettings, type CencoriProvider, type CencoriProviderSettings, cencori, createCencori };

package/dist/index.js

@@ -0,0 +1,305 @@
+"use strict";
+var __defProp = Object.defineProperty;
+var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
+var __getOwnPropNames = Object.getOwnPropertyNames;
+var __hasOwnProp = Object.prototype.hasOwnProperty;
+var __export = (target, all) => {
+  for (var name in all)
+    __defProp(target, name, { get: all[name], enumerable: true });
+};
+var __copyProps = (to, from, except, desc) => {
+  if (from && typeof from === "object" || typeof from === "function") {
+    for (let key of __getOwnPropNames(from))
+      if (!__hasOwnProp.call(to, key) && key !== except)
+        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
+  }
+  return to;
+};
+var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
+
+// src/index.ts
+var index_exports = {};
+__export(index_exports, {
+  CencoriChatLanguageModel: () => CencoriChatLanguageModel,
+  cencori: () => cencori,
+  createCencori: () => createCencori
+});
+module.exports = __toCommonJS(index_exports);
+
+// src/cencori-chat-model.ts
+var CencoriChatLanguageModel = class {
+  constructor(modelId, settings) {
+    this.specificationVersion = "v3";
+    this.provider = "cencori";
+    this.supportedUrls = {};
+    this.modelId = modelId;
+    this.settings = settings;
+  }
+  getHeaders() {
+    return {
+      "Content-Type": "application/json",
+      "CENCORI_API_KEY": this.settings.apiKey,
+      ...this.settings.headers
+    };
+  }
+  convertMessages(options) {
+    const messages = [];
+    const promptMessages = options.prompt;
+    if (!promptMessages || !Array.isArray(promptMessages)) {
+      return messages;
+    }
+    for (const msg of promptMessages) {
+      let content = "";
+      if (msg.role === "system") {
+        content = msg.content;
+      } else if (msg.role === "user" || msg.role === "assistant") {
+        const msgContent = msg.content;
+        if (Array.isArray(msgContent)) {
+          content = msgContent.filter((part) => part.type === "text").map((part) => part.text || "").join("");
+        } else if (typeof msgContent === "string") {
+          content = msgContent;
+        }
+      }
+      if (content && (msg.role === "system" || msg.role === "user" || msg.role === "assistant")) {
+        messages.push({
+          role: msg.role,
+          content
+        });
+      }
+    }
+    return messages;
+  }
+  mapFinishReason(reason) {
+    let unified;
+    switch (reason) {
+      case "stop":
+      case "end_turn":
+        unified = "stop";
+        break;
+      case "length":
+      case "max_tokens":
+        unified = "length";
+        break;
+      case "content_filter":
+        unified = "content-filter";
+        break;
+      case "tool_calls":
+      case "tool-calls":
+        unified = "tool-calls";
+        break;
+      case "error":
+        unified = "error";
+        break;
+      default:
+        unified = "stop";
+    }
+    return { unified, raw: reason };
+  }
+  buildUsage(inputTokens, outputTokens) {
+    return {
+      inputTokens: {
+        total: inputTokens,
+        noCache: inputTokens,
+        cacheRead: void 0,
+        cacheWrite: void 0
+      },
+      outputTokens: {
+        total: outputTokens,
+        text: outputTokens,
+        reasoning: void 0
+      }
+    };
+  }
+  async doGenerate(options) {
+    const messages = this.convertMessages(options);
+    const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {
+      method: "POST",
+      headers: this.getHeaders(),
+      body: JSON.stringify({
+        messages,
+        model: this.modelId,
+        temperature: options.temperature,
+        maxTokens: options.maxOutputTokens,
+        stream: false,
+        userId: this.settings.userId
+      }),
+      signal: options.abortSignal
+    });
+    if (!response.ok) {
+      const error = await response.json().catch(() => ({ error: "Unknown error" }));
+      throw new Error(`Cencori API error: ${error.error || response.statusText}`);
+    }
+    const data = await response.json();
+    const content = [{
+      type: "text",
+      text: data.content,
+      providerMetadata: void 0
+    }];
+    const warnings = [];
+    return {
+      content,
+      finishReason: this.mapFinishReason(data.finish_reason),
+      usage: this.buildUsage(data.usage.prompt_tokens, data.usage.completion_tokens),
+      warnings
+    };
+  }
+  async doStream(options) {
+    const messages = this.convertMessages(options);
+    const self = this;
+    const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {
+      method: "POST",
+      headers: this.getHeaders(),
+      body: JSON.stringify({
+        messages,
+        model: this.modelId,
+        temperature: options.temperature,
+        maxTokens: options.maxOutputTokens,
+        stream: true,
+        userId: this.settings.userId
+      }),
+      signal: options.abortSignal
+    });
+    if (!response.ok) {
+      const error = await response.json().catch(() => ({ error: "Unknown error" }));
+      throw new Error(`Cencori API error: ${error.error || response.statusText}`);
+    }
+    const reader = response.body?.getReader();
+    if (!reader) {
+      throw new Error("Response body is null");
+    }
+    const decoder = new TextDecoder();
+    let buffer = "";
+    let inputTokens = 0;
+    let outputTokens = 0;
+    const textPartId = "text-0";
+    let started = false;
+    const stream = new ReadableStream({
+      async pull(controller) {
+        try {
+          const { done, value } = await reader.read();
+          if (done) {
+            if (started) {
+              controller.enqueue({
+                type: "text-end",
+                id: textPartId
+              });
+            }
+            controller.enqueue({
+              type: "finish",
+              finishReason: self.mapFinishReason("stop"),
+              usage: self.buildUsage(inputTokens, outputTokens)
+            });
+            controller.close();
+            return;
+          }
+          buffer += decoder.decode(value, { stream: true });
+          const lines = buffer.split("\n");
+          buffer = lines.pop() || "";
+          for (const line of lines) {
+            if (line.trim() === "") continue;
+            if (!line.startsWith("data: ")) continue;
+            const data = line.slice(6);
+            if (data === "[DONE]") {
+              if (started) {
+                controller.enqueue({
+                  type: "text-end",
+                  id: textPartId
+                });
+              }
+              controller.enqueue({
+                type: "finish",
+                finishReason: self.mapFinishReason("stop"),
+                usage: self.buildUsage(inputTokens, outputTokens)
+              });
+              controller.close();
+              return;
+            }
+            try {
+              const chunk = JSON.parse(data);
+              if (chunk.delta) {
+                if (!started) {
+                  started = true;
+                  controller.enqueue({
+                    type: "text-start",
+                    id: textPartId
+                  });
+                }
+                outputTokens += Math.ceil(chunk.delta.length / 4);
+                controller.enqueue({
+                  type: "text-delta",
+                  id: textPartId,
+                  delta: chunk.delta
+                });
+              }
+              if (chunk.finish_reason) {
+                if (started) {
+                  controller.enqueue({
+                    type: "text-end",
+                    id: textPartId
+                  });
+                }
+                controller.enqueue({
+                  type: "finish",
+                  finishReason: self.mapFinishReason(chunk.finish_reason),
+                  usage: self.buildUsage(inputTokens, outputTokens)
+                });
+                controller.close();
+                return;
+              }
+            } catch {
+            }
+          }
+        } catch (error) {
+          controller.error(error);
+        }
+      },
+      cancel() {
+        reader.cancel();
+      }
+    });
+    return {
+      stream
+    };
+  }
+};
+
+// src/cencori-provider.ts
+function createCencori(options = {}) {
+  const baseUrl = options.baseUrl ?? "https://cencori.com";
+  const apiKey = options.apiKey ?? process.env.CENCORI_API_KEY;
+  if (!apiKey) {
+    throw new Error("Cencori API key is required. Pass it via options.apiKey or set CENCORI_API_KEY environment variable.");
+  }
+  const createModel = (modelId, settings = {}) => {
+    return new CencoriChatLanguageModel(modelId, {
+      apiKey,
+      baseUrl,
+      headers: options.headers,
+      ...settings
+    });
+  };
+  const provider = function(modelId, settings) {
+    return createModel(modelId, settings);
+  };
+  provider.chat = createModel;
+  return provider;
+}
+var cencori = function(modelId, settings) {
+  const apiKey = process.env.CENCORI_API_KEY;
+  if (!apiKey) {
+    throw new Error('CENCORI_API_KEY environment variable is required. Set it or use createCencori({ apiKey: "..." }) instead.');
+  }
+  return new CencoriChatLanguageModel(modelId, {
+    apiKey,
+    baseUrl: "https://cencori.com",
+    ...settings
+  });
+};
+cencori.chat = cencori;
+// Annotate the CommonJS export names for ESM import in node:
+0 && (module.exports = {
+  CencoriChatLanguageModel,
+  cencori,
+  createCencori
+});
+//# sourceMappingURL=index.js.map

package/dist/index.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/index.ts","../src/cencori-chat-model.ts","../src/cencori-provider.ts"],"sourcesContent":["/**\n * Cencori AI Provider for Vercel AI SDK\n * \n * @example\n * import { cencori } from '@cencori/ai-sdk';\n * import { streamText } from 'ai';\n * \n * const result = await streamText({\n *   model: cencori('gemini-2.5-flash'),\n *   messages: [{ role: 'user', content: 'Hello!' }]\n * });\n */\n\nexport { cencori, createCencori } from './cencori-provider';\nexport type { CencoriProvider } from './cencori-provider';\nexport type { CencoriProviderSettings, CencoriChatSettings } from './types';\nexport { CencoriChatLanguageModel } from './cencori-chat-model';\n","/**\n * Cencori Chat Language Model\n * \n * Implements the Vercel AI SDK's LanguageModelV3 interface (AI SDK v6 compatible)\n */\n\nimport type {\n    LanguageModelV3,\n    LanguageModelV3CallOptions,\n    LanguageModelV3GenerateResult,\n    LanguageModelV3StreamResult,\n    LanguageModelV3StreamPart,\n    LanguageModelV3Content,\n    LanguageModelV3Usage,\n    LanguageModelV3FinishReason,\n    SharedV3Warning,\n} from '@ai-sdk/provider';\n\nexport interface CencoriChatModelSettings {\n    apiKey: string;\n    baseUrl: string;\n    headers?: Record<string, string>;\n    userId?: string;\n}\n\ninterface CencoriMessage {\n    role: 'system' | 'user' | 'assistant';\n    content: string;\n}\n\ninterface CencoriResponse {\n    content: string;\n    model: string;\n    provider: string;\n    usage: {\n        prompt_tokens: number;\n        completion_tokens: number;\n        total_tokens: number;\n    };\n    cost_usd: number;\n    finish_reason?: string;\n}\n\ninterface CencoriStreamChunk {\n    delta: string;\n    finish_reason?: string;\n}\n\nexport class CencoriChatLanguageModel implements LanguageModelV3 {\n    readonly specificationVersion = 'v3' as const;\n    readonly provider = 'cencori';\n\n    readonly modelId: string;\n    readonly supportedUrls: Record<string, RegExp[]> = {};\n    private readonly settings: CencoriChatModelSettings;\n\n    constructor(modelId: string, settings: CencoriChatModelSettings) {\n        this.modelId = modelId;\n        this.settings = settings;\n    }\n\n    private getHeaders(): Record<string, string> {\n        return {\n            'Content-Type': 'application/json',\n            'CENCORI_API_KEY': this.settings.apiKey,\n            ...this.settings.headers,\n        };\n    }\n\n    private convertMessages(options: LanguageModelV3CallOptions): CencoriMessage[] {\n        const messages: CencoriMessage[] = [];\n\n        // In V3, options.prompt is directly an array of LanguageModelV3Message\n        const promptMessages = options.prompt;\n\n        if (!promptMessages || !Array.isArray(promptMessages)) {\n            return messages;\n        }\n\n        for (const msg of promptMessages) {\n            let content = '';\n\n            if (msg.role === 'system') {\n                // System messages have content as string directly\n                content = msg.content as string;\n            } else if (msg.role === 'user' || msg.role === 'assistant') {\n                // User and assistant messages have content as array of parts\n                const msgContent = msg.content;\n                if (Array.isArray(msgContent)) {\n                    content = msgContent\n                        .filter((part: { type: string }) => part.type === 'text')\n                        .map((part: { type: string; text?: string }) => part.text || '')\n                        .join('');\n                } else if (typeof msgContent === 'string') {\n                    content = msgContent;\n                }\n            }\n\n            if (content && (msg.role === 'system' || msg.role === 'user' || msg.role === 'assistant')) {\n                messages.push({\n                    role: msg.role as 'system' | 'user' | 'assistant',\n                    content,\n                });\n            }\n        }\n\n        return messages;\n    }\n\n    private mapFinishReason(reason?: string): LanguageModelV3FinishReason {\n        let unified: 'stop' | 'length' | 'content-filter' | 'tool-calls' | 'error' | 'other';\n\n        switch (reason) {\n            case 'stop':\n            case 'end_turn':\n                unified = 'stop';\n                break;\n            case 'length':\n            case 'max_tokens':\n                unified = 'length';\n                break;\n            case 'content_filter':\n                unified = 'content-filter';\n                break;\n            case 'tool_calls':\n            case 'tool-calls':\n                unified = 'tool-calls';\n                break;\n            case 'error':\n                unified = 'error';\n                break;\n            default:\n                unified = 'stop';\n        }\n\n        return { unified, raw: reason };\n    }\n\n    private buildUsage(inputTokens: number, outputTokens: number): LanguageModelV3Usage {\n        return {\n            inputTokens: {\n                total: inputTokens,\n                noCache: inputTokens,\n                cacheRead: undefined,\n                cacheWrite: undefined,\n            },\n            outputTokens: {\n                total: outputTokens,\n                text: outputTokens,\n                reasoning: undefined,\n            },\n        };\n    }\n\n    async doGenerate(options: LanguageModelV3CallOptions): Promise<LanguageModelV3GenerateResult> {\n        const messages = this.convertMessages(options);\n\n        const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {\n            method: 'POST',\n            headers: this.getHeaders(),\n            body: JSON.stringify({\n                messages,\n                model: this.modelId,\n                temperature: options.temperature,\n                maxTokens: options.maxOutputTokens,\n                stream: false,\n                userId: this.settings.userId,\n            }),\n            signal: options.abortSignal,\n        });\n\n        if (!response.ok) {\n            const error = await response.json().catch(() => ({ error: 'Unknown error' })) as { error?: string };\n            throw new Error(`Cencori API error: ${error.error || response.statusText}`);\n        }\n\n        const data = await response.json() as CencoriResponse;\n\n        const content: LanguageModelV3Content[] = [{\n            type: 'text',\n            text: data.content,\n            providerMetadata: undefined,\n        }];\n\n        const warnings: SharedV3Warning[] = [];\n\n        return {\n            content,\n            finishReason: this.mapFinishReason(data.finish_reason),\n            usage: this.buildUsage(data.usage.prompt_tokens, data.usage.completion_tokens),\n            warnings,\n        };\n    }\n\n    async doStream(options: LanguageModelV3CallOptions): Promise<LanguageModelV3StreamResult> {\n        const messages = this.convertMessages(options);\n        const self = this;\n\n        const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {\n            method: 'POST',\n            headers: this.getHeaders(),\n            body: JSON.stringify({\n                messages,\n                model: this.modelId,\n                temperature: options.temperature,\n                maxTokens: options.maxOutputTokens,\n                stream: true,\n                userId: this.settings.userId,\n            }),\n            signal: options.abortSignal,\n        });\n\n        if (!response.ok) {\n            const error = await response.json().catch(() => ({ error: 'Unknown error' })) as { error?: string };\n            throw new Error(`Cencori API error: ${error.error || response.statusText}`);\n        }\n\n        const reader = response.body?.getReader();\n        if (!reader) {\n            throw new Error('Response body is null');\n        }\n\n        const decoder = new TextDecoder();\n        let buffer = '';\n        let inputTokens = 0;\n        let outputTokens = 0;\n        const textPartId = 'text-0';\n        let started = false;\n\n        const stream = new ReadableStream<LanguageModelV3StreamPart>({\n            async pull(controller) {\n                try {\n                    const { done, value } = await reader.read();\n\n                    if (done) {\n                        // End text block and finish\n                        if (started) {\n                            controller.enqueue({\n                                type: 'text-end',\n                                id: textPartId,\n                            });\n                        }\n                        controller.enqueue({\n                            type: 'finish',\n                            finishReason: self.mapFinishReason('stop'),\n                            usage: self.buildUsage(inputTokens, outputTokens),\n                        });\n                        controller.close();\n                        return;\n                    }\n\n                    buffer += decoder.decode(value, { stream: true });\n                    const lines = buffer.split('\\n');\n                    buffer = lines.pop() || '';\n\n                    for (const line of lines) {\n                        if (line.trim() === '') continue;\n                        if (!line.startsWith('data: ')) continue;\n\n                        const data = line.slice(6);\n                        if (data === '[DONE]') {\n                            if (started) {\n                                controller.enqueue({\n                                    type: 'text-end',\n                                    id: textPartId,\n                                });\n                            }\n                            controller.enqueue({\n                                type: 'finish',\n                                finishReason: self.mapFinishReason('stop'),\n                                usage: self.buildUsage(inputTokens, outputTokens),\n                            });\n                            controller.close();\n                            return;\n                        }\n\n                        try {\n                            const chunk = JSON.parse(data) as CencoriStreamChunk;\n\n                            if (chunk.delta) {\n                                // Start text if not started\n                                if (!started) {\n                                    started = true;\n                                    controller.enqueue({\n                                        type: 'text-start',\n                                        id: textPartId,\n                                    });\n                                }\n\n                                outputTokens += Math.ceil(chunk.delta.length / 4); // Rough estimate\n                                controller.enqueue({\n                                    type: 'text-delta',\n                                    id: textPartId,\n                                    delta: chunk.delta,\n                                });\n                            }\n\n                            if (chunk.finish_reason) {\n                                if (started) {\n                                    controller.enqueue({\n                                        type: 'text-end',\n                                        id: textPartId,\n                                    });\n                                }\n                                controller.enqueue({\n                                    type: 'finish',\n                                    finishReason: self.mapFinishReason(chunk.finish_reason),\n                                    usage: self.buildUsage(inputTokens, outputTokens),\n                                });\n                                controller.close();\n                                return;\n                            }\n                        } catch {\n                            // Skip malformed JSON\n                        }\n                    }\n                } catch (error) {\n                    controller.error(error);\n                }\n            },\n            cancel() {\n                reader.cancel();\n            },\n        });\n\n        return {\n            stream,\n        };\n    }\n}\n","/**\n * Cencori AI Provider for Vercel AI SDK\n * \n * Use Cencori with streamText(), generateText(), and useChat()\n */\n\nimport { CencoriChatLanguageModel } from './cencori-chat-model';\nimport type { CencoriProviderSettings, CencoriChatSettings } from './types';\n\nexport interface CencoriProvider {\n    /**\n     * Create a Cencori chat model for use with Vercel AI SDK\n     * \n     * @param modelId - The model ID (e.g., 'gemini-2.5-flash', 'gpt-4o', 'claude-3-opus')\n     * @param settings - Optional model-specific settings\n     * @returns A LanguageModelV1 compatible model\n     * \n     * @example\n     * import { cencori } from '@cencori/ai-sdk';\n     * import { streamText } from 'ai';\n     * \n     * const result = await streamText({\n     *   model: cencori('gemini-2.5-flash'),\n     *   messages: [{ role: 'user', content: 'Hello!' }]\n     * });\n     */\n    (modelId: string, settings?: CencoriChatSettings): CencoriChatLanguageModel;\n\n    /**\n     * Create a chat model (alias for the provider function)\n     */\n    chat: (modelId: string, settings?: CencoriChatSettings) => CencoriChatLanguageModel;\n}\n\n/**\n * Create a Cencori provider instance\n * \n * @param options - Provider configuration options\n * @returns A Cencori provider\n * \n * @example\n * import { createCencori } from '@cencori/ai-sdk';\n * \n * const cencori = createCencori({\n *   apiKey: process.env.CENCORI_API_KEY\n * });\n * \n * const result = await streamText({\n *   model: cencori('gemini-2.5-flash'),\n *   messages: [{ role: 'user', content: 'Hello!' }]\n * });\n */\nexport function createCencori(options: CencoriProviderSettings = {}): CencoriProvider {\n    const baseUrl = options.baseUrl ?? 'https://cencori.com';\n    const apiKey = options.apiKey ?? process.env.CENCORI_API_KEY;\n\n    if (!apiKey) {\n        throw new Error('Cencori API key is required. Pass it via options.apiKey or set CENCORI_API_KEY environment variable.');\n    }\n\n    const createModel = (modelId: string, settings: CencoriChatSettings = {}) => {\n        return new CencoriChatLanguageModel(modelId, {\n            apiKey,\n            baseUrl,\n            headers: options.headers,\n            ...settings,\n        });\n    };\n\n    const provider = function (modelId: string, settings?: CencoriChatSettings) {\n        return createModel(modelId, settings);\n    } as CencoriProvider;\n\n    provider.chat = createModel;\n\n    return provider;\n}\n\n/**\n * Default Cencori provider instance\n * Uses CENCORI_API_KEY environment variable (lazy initialization)\n * \n * @example\n * import { cencori } from '@cencori/ai-sdk';\n * import { streamText } from 'ai';\n * \n * const result = await streamText({\n *   model: cencori('gemini-2.5-flash'),\n *   messages: [{ role: 'user', content: 'Hello!' }]\n * });\n */\nexport const cencori: CencoriProvider = function (modelId: string, settings?: CencoriChatSettings) {\n    const apiKey = process.env.CENCORI_API_KEY;\n    if (!apiKey) {\n        throw new Error('CENCORI_API_KEY environment variable is required. Set it or use createCencori({ apiKey: \"...\" }) instead.');\n    }\n    return new CencoriChatLanguageModel(modelId, {\n        apiKey,\n        baseUrl: 'https://cencori.com',\n        ...settings,\n    });\n} as CencoriProvider;\n\ncencori.chat = cencori;\n"],"mappings":";;;;;;;;;;;;;;;;;;;;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;AAAA;;;ACgDO,IAAM,2BAAN,MAA0D;AAAA,EAQ7D,YAAY,SAAiB,UAAoC;AAPjE,SAAS,uBAAuB;AAChC,SAAS,WAAW;AAGpB,SAAS,gBAA0C,CAAC;AAIhD,SAAK,UAAU;AACf,SAAK,WAAW;AAAA,EACpB;AAAA,EAEQ,aAAqC;AACzC,WAAO;AAAA,MACH,gBAAgB;AAAA,MAChB,mBAAmB,KAAK,SAAS;AAAA,MACjC,GAAG,KAAK,SAAS;AAAA,IACrB;AAAA,EACJ;AAAA,EAEQ,gBAAgB,SAAuD;AAC3E,UAAM,WAA6B,CAAC;AAGpC,UAAM,iBAAiB,QAAQ;AAE/B,QAAI,CAAC,kBAAkB,CAAC,MAAM,QAAQ,cAAc,GAAG;AACnD,aAAO;AAAA,IACX;AAEA,eAAW,OAAO,gBAAgB;AAC9B,UAAI,UAAU;AAEd,UAAI,IAAI,SAAS,UAAU;AAEvB,kBAAU,IAAI;AAAA,MAClB,WAAW,IAAI,SAAS,UAAU,IAAI,SAAS,aAAa;AAExD,cAAM,aAAa,IAAI;AACvB,YAAI,MAAM,QAAQ,UAAU,GAAG;AAC3B,oBAAU,WACL,OAAO,CAAC,SAA2B,KAAK,SAAS,MAAM,EACvD,IAAI,CAAC,SAA0C,KAAK,QAAQ,EAAE,EAC9D,KAAK,EAAE;AAAA,QAChB,WAAW,OAAO,eAAe,UAAU;AACvC,oBAAU;AAAA,QACd;AAAA,MACJ;AAEA,UAAI,YAAY,IAAI,SAAS,YAAY,IAAI,SAAS,UAAU,IAAI,SAAS,cAAc;AACvF,iBAAS,KAAK;AAAA,UACV,MAAM,IAAI;AAAA,UACV;AAAA,QACJ,CAAC;AAAA,MACL;AAAA,IACJ;AAEA,WAAO;AAAA,EACX;AAAA,EAEQ,gBAAgB,QAA8C;AAClE,QAAI;AAEJ,YAAQ,QAAQ;AAAA,MACZ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AACD,kBAAU;AACV;AAAA,MACJ;AACI,kBAAU;AAAA,IAClB;AAEA,WAAO,EAAE,SAAS,KAAK,OAAO;AAAA,EAClC;AAAA,EAEQ,WAAW,aAAqB,cAA4C;AAChF,WAAO;AAAA,MACH,aAAa;AAAA,QACT,OAAO;AAAA,QACP,SAAS;AAAA,QACT,WAAW;AAAA,QACX,YAAY;AAAA,MAChB;AAAA,MACA,cAAc;AAAA,QACV,OAAO;AAAA,QACP,MAAM;AAAA,QACN,WAAW;AAAA,MACf;AAAA,IACJ;AAAA,EACJ;AAAA,EAEA,MAAM,WAAW,SAA6E;AAC1F,UAAM,WAAW,KAAK,gBAAgB,OAAO;AAE7C,UAAM,WAAW,MAAM,MAAM,GAAG,KAAK,SAAS,OAAO,gBAAgB;AAAA,MACjE,QAAQ;AAAA,MACR,SAAS,KAAK,WAAW;AAAA,MACzB,MAAM,KAAK,UAAU;AAAA,QACjB;AAAA,QACA,OAAO,KAAK;AAAA,QACZ,aAAa,QAAQ;AAAA,QACrB,WAAW,QAAQ;AAAA,QACnB,QAAQ;AAAA,QACR,QAAQ,KAAK,SAAS;AAAA,MAC1B,CAAC;AAAA,MACD,QAAQ,QAAQ;AAAA,IACpB,CAAC;AAED,QAAI,CAAC,SAAS,IAAI;AACd,YAAM,QAAQ,MAAM,SAAS,KAAK,EAAE,MAAM,OAAO,EAAE,OAAO,gBAAgB,EAAE;AAC5E,YAAM,IAAI,MAAM,sBAAsB,MAAM,SAAS,SAAS,UAAU,EAAE;AAAA,IAC9E;AAEA,UAAM,OAAO,MAAM,SAAS,KAAK;AAEjC,UAAM,UAAoC,CAAC;AAAA,MACvC,MAAM;AAAA,MACN,MAAM,KAAK;AAAA,MACX,kBAAkB;AAAA,IACtB,CAAC;AAED,UAAM,WAA8B,CAAC;AAErC,WAAO;AAAA,MACH;AAAA,MACA,cAAc,KAAK,gBAAgB,KAAK,aAAa;AAAA,MACrD,OAAO,KAAK,WAAW,KAAK,MAAM,eAAe,KAAK,MAAM,iBAAiB;AAAA,MAC7E;AAAA,IACJ;AAAA,EACJ;AAAA,EAEA,MAAM,SAAS,SAA2E;AACtF,UAAM,WAAW,KAAK,gBAAgB,OAAO;AAC7C,UAAM,OAAO;AAEb,UAAM,WAAW,MAAM,MAAM,GAAG,KAAK,SAAS,OAAO,gBAAgB;AAAA,MACjE,QAAQ;AAAA,MACR,SAAS,KAAK,WAAW;AAAA,MACzB,MAAM,KAAK,UAAU;AAAA,QACjB;AAAA,QACA,OAAO,KAAK;AAAA,QACZ,aAAa,QAAQ;AAAA,QACrB,WAAW,QAAQ;AAAA,QACnB,QAAQ;AAAA,QACR,QAAQ,KAAK,SAAS;AAAA,MAC1B,CAAC;AAAA,MACD,QAAQ,QAAQ;AAAA,IACpB,CAAC;AAED,QAAI,CAAC,SAAS,IAAI;AACd,YAAM,QAAQ,MAAM,SAAS,KAAK,EAAE,MAAM,OAAO,EAAE,OAAO,gBAAgB,EAAE;AAC5E,YAAM,IAAI,MAAM,sBAAsB,MAAM,SAAS,SAAS,UAAU,EAAE;AAAA,IAC9E;AAEA,UAAM,SAAS,SAAS,MAAM,UAAU;AACxC,QAAI,CAAC,QAAQ;AACT,YAAM,IAAI,MAAM,uBAAuB;AAAA,IAC3C;AAEA,UAAM,UAAU,IAAI,YAAY;AAChC,QAAI,SAAS;AACb,QAAI,cAAc;AAClB,QAAI,eAAe;AACnB,UAAM,aAAa;AACnB,QAAI,UAAU;AAEd,UAAM,SAAS,IAAI,eAA0C;AAAA,MACzD,MAAM,KAAK,YAAY;AACnB,YAAI;AACA,gBAAM,EAAE,MAAM,MAAM,IAAI,MAAM,OAAO,KAAK;AAE1C,cAAI,MAAM;AAEN,gBAAI,SAAS;AACT,yBAAW,QAAQ;AAAA,gBACf,MAAM;AAAA,gBACN,IAAI;AAAA,cACR,CAAC;AAAA,YACL;AACA,uBAAW,QAAQ;AAAA,cACf,MAAM;AAAA,cACN,cAAc,KAAK,gBAAgB,MAAM;AAAA,cACzC,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,YACpD,CAAC;AACD,uBAAW,MAAM;AACjB;AAAA,UACJ;AAEA,oBAAU,QAAQ,OAAO,OAAO,EAAE,QAAQ,KAAK,CAAC;AAChD,gBAAM,QAAQ,OAAO,MAAM,IAAI;AAC/B,mBAAS,MAAM,IAAI,KAAK;AAExB,qBAAW,QAAQ,OAAO;AACtB,gBAAI,KAAK,KAAK,MAAM,GAAI;AACxB,gBAAI,CAAC,KAAK,WAAW,QAAQ,EAAG;AAEhC,kBAAM,OAAO,KAAK,MAAM,CAAC;AACzB,gBAAI,SAAS,UAAU;AACnB,kBAAI,SAAS;AACT,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,IAAI;AAAA,gBACR,CAAC;AAAA,cACL;AACA,yBAAW,QAAQ;AAAA,gBACf,MAAM;AAAA,gBACN,cAAc,KAAK,gBAAgB,MAAM;AAAA,gBACzC,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,cACpD,CAAC;AACD,yBAAW,MAAM;AACjB;AAAA,YACJ;AAEA,gBAAI;AACA,oBAAM,QAAQ,KAAK,MAAM,IAAI;AAE7B,kBAAI,MAAM,OAAO;AAEb,oBAAI,CAAC,SAAS;AACV,4BAAU;AACV,6BAAW,QAAQ;AAAA,oBACf,MAAM;AAAA,oBACN,IAAI;AAAA,kBACR,CAAC;AAAA,gBACL;AAEA,gCAAgB,KAAK,KAAK,MAAM,MAAM,SAAS,CAAC;AAChD,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,IAAI;AAAA,kBACJ,OAAO,MAAM;AAAA,gBACjB,CAAC;AAAA,cACL;AAEA,kBAAI,MAAM,eAAe;AACrB,oBAAI,SAAS;AACT,6BAAW,QAAQ;AAAA,oBACf,MAAM;AAAA,oBACN,IAAI;AAAA,kBACR,CAAC;AAAA,gBACL;AACA,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,cAAc,KAAK,gBAAgB,MAAM,aAAa;AAAA,kBACtD,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,gBACpD,CAAC;AACD,2BAAW,MAAM;AACjB;AAAA,cACJ;AAAA,YACJ,QAAQ;AAAA,YAER;AAAA,UACJ;AAAA,QACJ,SAAS,OAAO;AACZ,qBAAW,MAAM,KAAK;AAAA,QAC1B;AAAA,MACJ;AAAA,MACA,SAAS;AACL,eAAO,OAAO;AAAA,MAClB;AAAA,IACJ,CAAC;AAED,WAAO;AAAA,MACH;AAAA,IACJ;AAAA,EACJ;AACJ;;;ACrRO,SAAS,cAAc,UAAmC,CAAC,GAAoB;AAClF,QAAM,UAAU,QAAQ,WAAW;AACnC,QAAM,SAAS,QAAQ,UAAU,QAAQ,IAAI;AAE7C,MAAI,CAAC,QAAQ;AACT,UAAM,IAAI,MAAM,sGAAsG;AAAA,EAC1H;AAEA,QAAM,cAAc,CAAC,SAAiB,WAAgC,CAAC,MAAM;AACzE,WAAO,IAAI,yBAAyB,SAAS;AAAA,MACzC;AAAA,MACA;AAAA,MACA,SAAS,QAAQ;AAAA,MACjB,GAAG;AAAA,IACP,CAAC;AAAA,EACL;AAEA,QAAM,WAAW,SAAU,SAAiB,UAAgC;AACxE,WAAO,YAAY,SAAS,QAAQ;AAAA,EACxC;AAEA,WAAS,OAAO;AAEhB,SAAO;AACX;AAeO,IAAM,UAA2B,SAAU,SAAiB,UAAgC;AAC/F,QAAM,SAAS,QAAQ,IAAI;AAC3B,MAAI,CAAC,QAAQ;AACT,UAAM,IAAI,MAAM,2GAA2G;AAAA,EAC/H;AACA,SAAO,IAAI,yBAAyB,SAAS;AAAA,IACzC;AAAA,IACA,SAAS;AAAA,IACT,GAAG;AAAA,EACP,CAAC;AACL;AAEA,QAAQ,OAAO;","names":[]}

package/dist/index.mjs

@@ -0,0 +1,276 @@
+// src/cencori-chat-model.ts
+var CencoriChatLanguageModel = class {
+  constructor(modelId, settings) {
+    this.specificationVersion = "v3";
+    this.provider = "cencori";
+    this.supportedUrls = {};
+    this.modelId = modelId;
+    this.settings = settings;
+  }
+  getHeaders() {
+    return {
+      "Content-Type": "application/json",
+      "CENCORI_API_KEY": this.settings.apiKey,
+      ...this.settings.headers
+    };
+  }
+  convertMessages(options) {
+    const messages = [];
+    const promptMessages = options.prompt;
+    if (!promptMessages || !Array.isArray(promptMessages)) {
+      return messages;
+    }
+    for (const msg of promptMessages) {
+      let content = "";
+      if (msg.role === "system") {
+        content = msg.content;
+      } else if (msg.role === "user" || msg.role === "assistant") {
+        const msgContent = msg.content;
+        if (Array.isArray(msgContent)) {
+          content = msgContent.filter((part) => part.type === "text").map((part) => part.text || "").join("");
+        } else if (typeof msgContent === "string") {
+          content = msgContent;
+        }
+      }
+      if (content && (msg.role === "system" || msg.role === "user" || msg.role === "assistant")) {
+        messages.push({
+          role: msg.role,
+          content
+        });
+      }
+    }
+    return messages;
+  }
+  mapFinishReason(reason) {
+    let unified;
+    switch (reason) {
+      case "stop":
+      case "end_turn":
+        unified = "stop";
+        break;
+      case "length":
+      case "max_tokens":
+        unified = "length";
+        break;
+      case "content_filter":
+        unified = "content-filter";
+        break;
+      case "tool_calls":
+      case "tool-calls":
+        unified = "tool-calls";
+        break;
+      case "error":
+        unified = "error";
+        break;
+      default:
+        unified = "stop";
+    }
+    return { unified, raw: reason };
+  }
+  buildUsage(inputTokens, outputTokens) {
+    return {
+      inputTokens: {
+        total: inputTokens,
+        noCache: inputTokens,
+        cacheRead: void 0,
+        cacheWrite: void 0
+      },
+      outputTokens: {
+        total: outputTokens,
+        text: outputTokens,
+        reasoning: void 0
+      }
+    };
+  }
+  async doGenerate(options) {
+    const messages = this.convertMessages(options);
+    const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {
+      method: "POST",
+      headers: this.getHeaders(),
+      body: JSON.stringify({
+        messages,
+        model: this.modelId,
+        temperature: options.temperature,
+        maxTokens: options.maxOutputTokens,
+        stream: false,
+        userId: this.settings.userId
+      }),
+      signal: options.abortSignal
+    });
+    if (!response.ok) {
+      const error = await response.json().catch(() => ({ error: "Unknown error" }));
+      throw new Error(`Cencori API error: ${error.error || response.statusText}`);
+    }
+    const data = await response.json();
+    const content = [{
+      type: "text",
+      text: data.content,
+      providerMetadata: void 0
+    }];
+    const warnings = [];
+    return {
+      content,
+      finishReason: this.mapFinishReason(data.finish_reason),
+      usage: this.buildUsage(data.usage.prompt_tokens, data.usage.completion_tokens),
+      warnings
+    };
+  }
+  async doStream(options) {
+    const messages = this.convertMessages(options);
+    const self = this;
+    const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {
+      method: "POST",
+      headers: this.getHeaders(),
+      body: JSON.stringify({
+        messages,
+        model: this.modelId,
+        temperature: options.temperature,
+        maxTokens: options.maxOutputTokens,
+        stream: true,
+        userId: this.settings.userId
+      }),
+      signal: options.abortSignal
+    });
+    if (!response.ok) {
+      const error = await response.json().catch(() => ({ error: "Unknown error" }));
+      throw new Error(`Cencori API error: ${error.error || response.statusText}`);
+    }
+    const reader = response.body?.getReader();
+    if (!reader) {
+      throw new Error("Response body is null");
+    }
+    const decoder = new TextDecoder();
+    let buffer = "";
+    let inputTokens = 0;
+    let outputTokens = 0;
+    const textPartId = "text-0";
+    let started = false;
+    const stream = new ReadableStream({
+      async pull(controller) {
+        try {
+          const { done, value } = await reader.read();
+          if (done) {
+            if (started) {
+              controller.enqueue({
+                type: "text-end",
+                id: textPartId
+              });
+            }
+            controller.enqueue({
+              type: "finish",
+              finishReason: self.mapFinishReason("stop"),
+              usage: self.buildUsage(inputTokens, outputTokens)
+            });
+            controller.close();
+            return;
+          }
+          buffer += decoder.decode(value, { stream: true });
+          const lines = buffer.split("\n");
+          buffer = lines.pop() || "";
+          for (const line of lines) {
+            if (line.trim() === "") continue;
+            if (!line.startsWith("data: ")) continue;
+            const data = line.slice(6);
+            if (data === "[DONE]") {
+              if (started) {
+                controller.enqueue({
+                  type: "text-end",
+                  id: textPartId
+                });
+              }
+              controller.enqueue({
+                type: "finish",
+                finishReason: self.mapFinishReason("stop"),
+                usage: self.buildUsage(inputTokens, outputTokens)
+              });
+              controller.close();
+              return;
+            }
+            try {
+              const chunk = JSON.parse(data);
+              if (chunk.delta) {
+                if (!started) {
+                  started = true;
+                  controller.enqueue({
+                    type: "text-start",
+                    id: textPartId
+                  });
+                }
+                outputTokens += Math.ceil(chunk.delta.length / 4);
+                controller.enqueue({
+                  type: "text-delta",
+                  id: textPartId,
+                  delta: chunk.delta
+                });
+              }
+              if (chunk.finish_reason) {
+                if (started) {
+                  controller.enqueue({
+                    type: "text-end",
+                    id: textPartId
+                  });
+                }
+                controller.enqueue({
+                  type: "finish",
+                  finishReason: self.mapFinishReason(chunk.finish_reason),
+                  usage: self.buildUsage(inputTokens, outputTokens)
+                });
+                controller.close();
+                return;
+              }
+            } catch {
+            }
+          }
+        } catch (error) {
+          controller.error(error);
+        }
+      },
+      cancel() {
+        reader.cancel();
+      }
+    });
+    return {
+      stream
+    };
+  }
+};
+
+// src/cencori-provider.ts
+function createCencori(options = {}) {
+  const baseUrl = options.baseUrl ?? "https://cencori.com";
+  const apiKey = options.apiKey ?? process.env.CENCORI_API_KEY;
+  if (!apiKey) {
+    throw new Error("Cencori API key is required. Pass it via options.apiKey or set CENCORI_API_KEY environment variable.");
+  }
+  const createModel = (modelId, settings = {}) => {
+    return new CencoriChatLanguageModel(modelId, {
+      apiKey,
+      baseUrl,
+      headers: options.headers,
+      ...settings
+    });
+  };
+  const provider = function(modelId, settings) {
+    return createModel(modelId, settings);
+  };
+  provider.chat = createModel;
+  return provider;
+}
+var cencori = function(modelId, settings) {
+  const apiKey = process.env.CENCORI_API_KEY;
+  if (!apiKey) {
+    throw new Error('CENCORI_API_KEY environment variable is required. Set it or use createCencori({ apiKey: "..." }) instead.');
+  }
+  return new CencoriChatLanguageModel(modelId, {
+    apiKey,
+    baseUrl: "https://cencori.com",
+    ...settings
+  });
+};
+cencori.chat = cencori;
+export {
+  CencoriChatLanguageModel,
+  cencori,
+  createCencori
+};
+//# sourceMappingURL=index.mjs.map

package/dist/index.mjs.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/cencori-chat-model.ts","../src/cencori-provider.ts"],"sourcesContent":["/**\n * Cencori Chat Language Model\n * \n * Implements the Vercel AI SDK's LanguageModelV3 interface (AI SDK v6 compatible)\n */\n\nimport type {\n    LanguageModelV3,\n    LanguageModelV3CallOptions,\n    LanguageModelV3GenerateResult,\n    LanguageModelV3StreamResult,\n    LanguageModelV3StreamPart,\n    LanguageModelV3Content,\n    LanguageModelV3Usage,\n    LanguageModelV3FinishReason,\n    SharedV3Warning,\n} from '@ai-sdk/provider';\n\nexport interface CencoriChatModelSettings {\n    apiKey: string;\n    baseUrl: string;\n    headers?: Record<string, string>;\n    userId?: string;\n}\n\ninterface CencoriMessage {\n    role: 'system' | 'user' | 'assistant';\n    content: string;\n}\n\ninterface CencoriResponse {\n    content: string;\n    model: string;\n    provider: string;\n    usage: {\n        prompt_tokens: number;\n        completion_tokens: number;\n        total_tokens: number;\n    };\n    cost_usd: number;\n    finish_reason?: string;\n}\n\ninterface CencoriStreamChunk {\n    delta: string;\n    finish_reason?: string;\n}\n\nexport class CencoriChatLanguageModel implements LanguageModelV3 {\n    readonly specificationVersion = 'v3' as const;\n    readonly provider = 'cencori';\n\n    readonly modelId: string;\n    readonly supportedUrls: Record<string, RegExp[]> = {};\n    private readonly settings: CencoriChatModelSettings;\n\n    constructor(modelId: string, settings: CencoriChatModelSettings) {\n        this.modelId = modelId;\n        this.settings = settings;\n    }\n\n    private getHeaders(): Record<string, string> {\n        return {\n            'Content-Type': 'application/json',\n            'CENCORI_API_KEY': this.settings.apiKey,\n            ...this.settings.headers,\n        };\n    }\n\n    private convertMessages(options: LanguageModelV3CallOptions): CencoriMessage[] {\n        const messages: CencoriMessage[] = [];\n\n        // In V3, options.prompt is directly an array of LanguageModelV3Message\n        const promptMessages = options.prompt;\n\n        if (!promptMessages || !Array.isArray(promptMessages)) {\n            return messages;\n        }\n\n        for (const msg of promptMessages) {\n            let content = '';\n\n            if (msg.role === 'system') {\n                // System messages have content as string directly\n                content = msg.content as string;\n            } else if (msg.role === 'user' || msg.role === 'assistant') {\n                // User and assistant messages have content as array of parts\n                const msgContent = msg.content;\n                if (Array.isArray(msgContent)) {\n                    content = msgContent\n                        .filter((part: { type: string }) => part.type === 'text')\n                        .map((part: { type: string; text?: string }) => part.text || '')\n                        .join('');\n                } else if (typeof msgContent === 'string') {\n                    content = msgContent;\n                }\n            }\n\n            if (content && (msg.role === 'system' || msg.role === 'user' || msg.role === 'assistant')) {\n                messages.push({\n                    role: msg.role as 'system' | 'user' | 'assistant',\n                    content,\n                });\n            }\n        }\n\n        return messages;\n    }\n\n    private mapFinishReason(reason?: string): LanguageModelV3FinishReason {\n        let unified: 'stop' | 'length' | 'content-filter' | 'tool-calls' | 'error' | 'other';\n\n        switch (reason) {\n            case 'stop':\n            case 'end_turn':\n                unified = 'stop';\n                break;\n            case 'length':\n            case 'max_tokens':\n                unified = 'length';\n                break;\n            case 'content_filter':\n                unified = 'content-filter';\n                break;\n            case 'tool_calls':\n            case 'tool-calls':\n                unified = 'tool-calls';\n                break;\n            case 'error':\n                unified = 'error';\n                break;\n            default:\n                unified = 'stop';\n        }\n\n        return { unified, raw: reason };\n    }\n\n    private buildUsage(inputTokens: number, outputTokens: number): LanguageModelV3Usage {\n        return {\n            inputTokens: {\n                total: inputTokens,\n                noCache: inputTokens,\n                cacheRead: undefined,\n                cacheWrite: undefined,\n            },\n            outputTokens: {\n                total: outputTokens,\n                text: outputTokens,\n                reasoning: undefined,\n            },\n        };\n    }\n\n    async doGenerate(options: LanguageModelV3CallOptions): Promise<LanguageModelV3GenerateResult> {\n        const messages = this.convertMessages(options);\n\n        const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {\n            method: 'POST',\n            headers: this.getHeaders(),\n            body: JSON.stringify({\n                messages,\n                model: this.modelId,\n                temperature: options.temperature,\n                maxTokens: options.maxOutputTokens,\n                stream: false,\n                userId: this.settings.userId,\n            }),\n            signal: options.abortSignal,\n        });\n\n        if (!response.ok) {\n            const error = await response.json().catch(() => ({ error: 'Unknown error' })) as { error?: string };\n            throw new Error(`Cencori API error: ${error.error || response.statusText}`);\n        }\n\n        const data = await response.json() as CencoriResponse;\n\n        const content: LanguageModelV3Content[] = [{\n            type: 'text',\n            text: data.content,\n            providerMetadata: undefined,\n        }];\n\n        const warnings: SharedV3Warning[] = [];\n\n        return {\n            content,\n            finishReason: this.mapFinishReason(data.finish_reason),\n            usage: this.buildUsage(data.usage.prompt_tokens, data.usage.completion_tokens),\n            warnings,\n        };\n    }\n\n    async doStream(options: LanguageModelV3CallOptions): Promise<LanguageModelV3StreamResult> {\n        const messages = this.convertMessages(options);\n        const self = this;\n\n        const response = await fetch(`${this.settings.baseUrl}/api/ai/chat`, {\n            method: 'POST',\n            headers: this.getHeaders(),\n            body: JSON.stringify({\n                messages,\n                model: this.modelId,\n                temperature: options.temperature,\n                maxTokens: options.maxOutputTokens,\n                stream: true,\n                userId: this.settings.userId,\n            }),\n            signal: options.abortSignal,\n        });\n\n        if (!response.ok) {\n            const error = await response.json().catch(() => ({ error: 'Unknown error' })) as { error?: string };\n            throw new Error(`Cencori API error: ${error.error || response.statusText}`);\n        }\n\n        const reader = response.body?.getReader();\n        if (!reader) {\n            throw new Error('Response body is null');\n        }\n\n        const decoder = new TextDecoder();\n        let buffer = '';\n        let inputTokens = 0;\n        let outputTokens = 0;\n        const textPartId = 'text-0';\n        let started = false;\n\n        const stream = new ReadableStream<LanguageModelV3StreamPart>({\n            async pull(controller) {\n                try {\n                    const { done, value } = await reader.read();\n\n                    if (done) {\n                        // End text block and finish\n                        if (started) {\n                            controller.enqueue({\n                                type: 'text-end',\n                                id: textPartId,\n                            });\n                        }\n                        controller.enqueue({\n                            type: 'finish',\n                            finishReason: self.mapFinishReason('stop'),\n                            usage: self.buildUsage(inputTokens, outputTokens),\n                        });\n                        controller.close();\n                        return;\n                    }\n\n                    buffer += decoder.decode(value, { stream: true });\n                    const lines = buffer.split('\\n');\n                    buffer = lines.pop() || '';\n\n                    for (const line of lines) {\n                        if (line.trim() === '') continue;\n                        if (!line.startsWith('data: ')) continue;\n\n                        const data = line.slice(6);\n                        if (data === '[DONE]') {\n                            if (started) {\n                                controller.enqueue({\n                                    type: 'text-end',\n                                    id: textPartId,\n                                });\n                            }\n                            controller.enqueue({\n                                type: 'finish',\n                                finishReason: self.mapFinishReason('stop'),\n                                usage: self.buildUsage(inputTokens, outputTokens),\n                            });\n                            controller.close();\n                            return;\n                        }\n\n                        try {\n                            const chunk = JSON.parse(data) as CencoriStreamChunk;\n\n                            if (chunk.delta) {\n                                // Start text if not started\n                                if (!started) {\n                                    started = true;\n                                    controller.enqueue({\n                                        type: 'text-start',\n                                        id: textPartId,\n                                    });\n                                }\n\n                                outputTokens += Math.ceil(chunk.delta.length / 4); // Rough estimate\n                                controller.enqueue({\n                                    type: 'text-delta',\n                                    id: textPartId,\n                                    delta: chunk.delta,\n                                });\n                            }\n\n                            if (chunk.finish_reason) {\n                                if (started) {\n                                    controller.enqueue({\n                                        type: 'text-end',\n                                        id: textPartId,\n                                    });\n                                }\n                                controller.enqueue({\n                                    type: 'finish',\n                                    finishReason: self.mapFinishReason(chunk.finish_reason),\n                                    usage: self.buildUsage(inputTokens, outputTokens),\n                                });\n                                controller.close();\n                                return;\n                            }\n                        } catch {\n                            // Skip malformed JSON\n                        }\n                    }\n                } catch (error) {\n                    controller.error(error);\n                }\n            },\n            cancel() {\n                reader.cancel();\n            },\n        });\n\n        return {\n            stream,\n        };\n    }\n}\n","/**\n * Cencori AI Provider for Vercel AI SDK\n * \n * Use Cencori with streamText(), generateText(), and useChat()\n */\n\nimport { CencoriChatLanguageModel } from './cencori-chat-model';\nimport type { CencoriProviderSettings, CencoriChatSettings } from './types';\n\nexport interface CencoriProvider {\n    /**\n     * Create a Cencori chat model for use with Vercel AI SDK\n     * \n     * @param modelId - The model ID (e.g., 'gemini-2.5-flash', 'gpt-4o', 'claude-3-opus')\n     * @param settings - Optional model-specific settings\n     * @returns A LanguageModelV1 compatible model\n     * \n     * @example\n     * import { cencori } from '@cencori/ai-sdk';\n     * import { streamText } from 'ai';\n     * \n     * const result = await streamText({\n     *   model: cencori('gemini-2.5-flash'),\n     *   messages: [{ role: 'user', content: 'Hello!' }]\n     * });\n     */\n    (modelId: string, settings?: CencoriChatSettings): CencoriChatLanguageModel;\n\n    /**\n     * Create a chat model (alias for the provider function)\n     */\n    chat: (modelId: string, settings?: CencoriChatSettings) => CencoriChatLanguageModel;\n}\n\n/**\n * Create a Cencori provider instance\n * \n * @param options - Provider configuration options\n * @returns A Cencori provider\n * \n * @example\n * import { createCencori } from '@cencori/ai-sdk';\n * \n * const cencori = createCencori({\n *   apiKey: process.env.CENCORI_API_KEY\n * });\n * \n * const result = await streamText({\n *   model: cencori('gemini-2.5-flash'),\n *   messages: [{ role: 'user', content: 'Hello!' }]\n * });\n */\nexport function createCencori(options: CencoriProviderSettings = {}): CencoriProvider {\n    const baseUrl = options.baseUrl ?? 'https://cencori.com';\n    const apiKey = options.apiKey ?? process.env.CENCORI_API_KEY;\n\n    if (!apiKey) {\n        throw new Error('Cencori API key is required. Pass it via options.apiKey or set CENCORI_API_KEY environment variable.');\n    }\n\n    const createModel = (modelId: string, settings: CencoriChatSettings = {}) => {\n        return new CencoriChatLanguageModel(modelId, {\n            apiKey,\n            baseUrl,\n            headers: options.headers,\n            ...settings,\n        });\n    };\n\n    const provider = function (modelId: string, settings?: CencoriChatSettings) {\n        return createModel(modelId, settings);\n    } as CencoriProvider;\n\n    provider.chat = createModel;\n\n    return provider;\n}\n\n/**\n * Default Cencori provider instance\n * Uses CENCORI_API_KEY environment variable (lazy initialization)\n * \n * @example\n * import { cencori } from '@cencori/ai-sdk';\n * import { streamText } from 'ai';\n * \n * const result = await streamText({\n *   model: cencori('gemini-2.5-flash'),\n *   messages: [{ role: 'user', content: 'Hello!' }]\n * });\n */\nexport const cencori: CencoriProvider = function (modelId: string, settings?: CencoriChatSettings) {\n    const apiKey = process.env.CENCORI_API_KEY;\n    if (!apiKey) {\n        throw new Error('CENCORI_API_KEY environment variable is required. Set it or use createCencori({ apiKey: \"...\" }) instead.');\n    }\n    return new CencoriChatLanguageModel(modelId, {\n        apiKey,\n        baseUrl: 'https://cencori.com',\n        ...settings,\n    });\n} as CencoriProvider;\n\ncencori.chat = cencori;\n"],"mappings":";AAgDO,IAAM,2BAAN,MAA0D;AAAA,EAQ7D,YAAY,SAAiB,UAAoC;AAPjE,SAAS,uBAAuB;AAChC,SAAS,WAAW;AAGpB,SAAS,gBAA0C,CAAC;AAIhD,SAAK,UAAU;AACf,SAAK,WAAW;AAAA,EACpB;AAAA,EAEQ,aAAqC;AACzC,WAAO;AAAA,MACH,gBAAgB;AAAA,MAChB,mBAAmB,KAAK,SAAS;AAAA,MACjC,GAAG,KAAK,SAAS;AAAA,IACrB;AAAA,EACJ;AAAA,EAEQ,gBAAgB,SAAuD;AAC3E,UAAM,WAA6B,CAAC;AAGpC,UAAM,iBAAiB,QAAQ;AAE/B,QAAI,CAAC,kBAAkB,CAAC,MAAM,QAAQ,cAAc,GAAG;AACnD,aAAO;AAAA,IACX;AAEA,eAAW,OAAO,gBAAgB;AAC9B,UAAI,UAAU;AAEd,UAAI,IAAI,SAAS,UAAU;AAEvB,kBAAU,IAAI;AAAA,MAClB,WAAW,IAAI,SAAS,UAAU,IAAI,SAAS,aAAa;AAExD,cAAM,aAAa,IAAI;AACvB,YAAI,MAAM,QAAQ,UAAU,GAAG;AAC3B,oBAAU,WACL,OAAO,CAAC,SAA2B,KAAK,SAAS,MAAM,EACvD,IAAI,CAAC,SAA0C,KAAK,QAAQ,EAAE,EAC9D,KAAK,EAAE;AAAA,QAChB,WAAW,OAAO,eAAe,UAAU;AACvC,oBAAU;AAAA,QACd;AAAA,MACJ;AAEA,UAAI,YAAY,IAAI,SAAS,YAAY,IAAI,SAAS,UAAU,IAAI,SAAS,cAAc;AACvF,iBAAS,KAAK;AAAA,UACV,MAAM,IAAI;AAAA,UACV;AAAA,QACJ,CAAC;AAAA,MACL;AAAA,IACJ;AAEA,WAAO;AAAA,EACX;AAAA,EAEQ,gBAAgB,QAA8C;AAClE,QAAI;AAEJ,YAAQ,QAAQ;AAAA,MACZ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AAAA,MACL,KAAK;AACD,kBAAU;AACV;AAAA,MACJ,KAAK;AACD,kBAAU;AACV;AAAA,MACJ;AACI,kBAAU;AAAA,IAClB;AAEA,WAAO,EAAE,SAAS,KAAK,OAAO;AAAA,EAClC;AAAA,EAEQ,WAAW,aAAqB,cAA4C;AAChF,WAAO;AAAA,MACH,aAAa;AAAA,QACT,OAAO;AAAA,QACP,SAAS;AAAA,QACT,WAAW;AAAA,QACX,YAAY;AAAA,MAChB;AAAA,MACA,cAAc;AAAA,QACV,OAAO;AAAA,QACP,MAAM;AAAA,QACN,WAAW;AAAA,MACf;AAAA,IACJ;AAAA,EACJ;AAAA,EAEA,MAAM,WAAW,SAA6E;AAC1F,UAAM,WAAW,KAAK,gBAAgB,OAAO;AAE7C,UAAM,WAAW,MAAM,MAAM,GAAG,KAAK,SAAS,OAAO,gBAAgB;AAAA,MACjE,QAAQ;AAAA,MACR,SAAS,KAAK,WAAW;AAAA,MACzB,MAAM,KAAK,UAAU;AAAA,QACjB;AAAA,QACA,OAAO,KAAK;AAAA,QACZ,aAAa,QAAQ;AAAA,QACrB,WAAW,QAAQ;AAAA,QACnB,QAAQ;AAAA,QACR,QAAQ,KAAK,SAAS;AAAA,MAC1B,CAAC;AAAA,MACD,QAAQ,QAAQ;AAAA,IACpB,CAAC;AAED,QAAI,CAAC,SAAS,IAAI;AACd,YAAM,QAAQ,MAAM,SAAS,KAAK,EAAE,MAAM,OAAO,EAAE,OAAO,gBAAgB,EAAE;AAC5E,YAAM,IAAI,MAAM,sBAAsB,MAAM,SAAS,SAAS,UAAU,EAAE;AAAA,IAC9E;AAEA,UAAM,OAAO,MAAM,SAAS,KAAK;AAEjC,UAAM,UAAoC,CAAC;AAAA,MACvC,MAAM;AAAA,MACN,MAAM,KAAK;AAAA,MACX,kBAAkB;AAAA,IACtB,CAAC;AAED,UAAM,WAA8B,CAAC;AAErC,WAAO;AAAA,MACH;AAAA,MACA,cAAc,KAAK,gBAAgB,KAAK,aAAa;AAAA,MACrD,OAAO,KAAK,WAAW,KAAK,MAAM,eAAe,KAAK,MAAM,iBAAiB;AAAA,MAC7E;AAAA,IACJ;AAAA,EACJ;AAAA,EAEA,MAAM,SAAS,SAA2E;AACtF,UAAM,WAAW,KAAK,gBAAgB,OAAO;AAC7C,UAAM,OAAO;AAEb,UAAM,WAAW,MAAM,MAAM,GAAG,KAAK,SAAS,OAAO,gBAAgB;AAAA,MACjE,QAAQ;AAAA,MACR,SAAS,KAAK,WAAW;AAAA,MACzB,MAAM,KAAK,UAAU;AAAA,QACjB;AAAA,QACA,OAAO,KAAK;AAAA,QACZ,aAAa,QAAQ;AAAA,QACrB,WAAW,QAAQ;AAAA,QACnB,QAAQ;AAAA,QACR,QAAQ,KAAK,SAAS;AAAA,MAC1B,CAAC;AAAA,MACD,QAAQ,QAAQ;AAAA,IACpB,CAAC;AAED,QAAI,CAAC,SAAS,IAAI;AACd,YAAM,QAAQ,MAAM,SAAS,KAAK,EAAE,MAAM,OAAO,EAAE,OAAO,gBAAgB,EAAE;AAC5E,YAAM,IAAI,MAAM,sBAAsB,MAAM,SAAS,SAAS,UAAU,EAAE;AAAA,IAC9E;AAEA,UAAM,SAAS,SAAS,MAAM,UAAU;AACxC,QAAI,CAAC,QAAQ;AACT,YAAM,IAAI,MAAM,uBAAuB;AAAA,IAC3C;AAEA,UAAM,UAAU,IAAI,YAAY;AAChC,QAAI,SAAS;AACb,QAAI,cAAc;AAClB,QAAI,eAAe;AACnB,UAAM,aAAa;AACnB,QAAI,UAAU;AAEd,UAAM,SAAS,IAAI,eAA0C;AAAA,MACzD,MAAM,KAAK,YAAY;AACnB,YAAI;AACA,gBAAM,EAAE,MAAM,MAAM,IAAI,MAAM,OAAO,KAAK;AAE1C,cAAI,MAAM;AAEN,gBAAI,SAAS;AACT,yBAAW,QAAQ;AAAA,gBACf,MAAM;AAAA,gBACN,IAAI;AAAA,cACR,CAAC;AAAA,YACL;AACA,uBAAW,QAAQ;AAAA,cACf,MAAM;AAAA,cACN,cAAc,KAAK,gBAAgB,MAAM;AAAA,cACzC,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,YACpD,CAAC;AACD,uBAAW,MAAM;AACjB;AAAA,UACJ;AAEA,oBAAU,QAAQ,OAAO,OAAO,EAAE,QAAQ,KAAK,CAAC;AAChD,gBAAM,QAAQ,OAAO,MAAM,IAAI;AAC/B,mBAAS,MAAM,IAAI,KAAK;AAExB,qBAAW,QAAQ,OAAO;AACtB,gBAAI,KAAK,KAAK,MAAM,GAAI;AACxB,gBAAI,CAAC,KAAK,WAAW,QAAQ,EAAG;AAEhC,kBAAM,OAAO,KAAK,MAAM,CAAC;AACzB,gBAAI,SAAS,UAAU;AACnB,kBAAI,SAAS;AACT,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,IAAI;AAAA,gBACR,CAAC;AAAA,cACL;AACA,yBAAW,QAAQ;AAAA,gBACf,MAAM;AAAA,gBACN,cAAc,KAAK,gBAAgB,MAAM;AAAA,gBACzC,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,cACpD,CAAC;AACD,yBAAW,MAAM;AACjB;AAAA,YACJ;AAEA,gBAAI;AACA,oBAAM,QAAQ,KAAK,MAAM,IAAI;AAE7B,kBAAI,MAAM,OAAO;AAEb,oBAAI,CAAC,SAAS;AACV,4BAAU;AACV,6BAAW,QAAQ;AAAA,oBACf,MAAM;AAAA,oBACN,IAAI;AAAA,kBACR,CAAC;AAAA,gBACL;AAEA,gCAAgB,KAAK,KAAK,MAAM,MAAM,SAAS,CAAC;AAChD,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,IAAI;AAAA,kBACJ,OAAO,MAAM;AAAA,gBACjB,CAAC;AAAA,cACL;AAEA,kBAAI,MAAM,eAAe;AACrB,oBAAI,SAAS;AACT,6BAAW,QAAQ;AAAA,oBACf,MAAM;AAAA,oBACN,IAAI;AAAA,kBACR,CAAC;AAAA,gBACL;AACA,2BAAW,QAAQ;AAAA,kBACf,MAAM;AAAA,kBACN,cAAc,KAAK,gBAAgB,MAAM,aAAa;AAAA,kBACtD,OAAO,KAAK,WAAW,aAAa,YAAY;AAAA,gBACpD,CAAC;AACD,2BAAW,MAAM;AACjB;AAAA,cACJ;AAAA,YACJ,QAAQ;AAAA,YAER;AAAA,UACJ;AAAA,QACJ,SAAS,OAAO;AACZ,qBAAW,MAAM,KAAK;AAAA,QAC1B;AAAA,MACJ;AAAA,MACA,SAAS;AACL,eAAO,OAAO;AAAA,MAClB;AAAA,IACJ,CAAC;AAED,WAAO;AAAA,MACH;AAAA,IACJ;AAAA,EACJ;AACJ;;;ACrRO,SAAS,cAAc,UAAmC,CAAC,GAAoB;AAClF,QAAM,UAAU,QAAQ,WAAW;AACnC,QAAM,SAAS,QAAQ,UAAU,QAAQ,IAAI;AAE7C,MAAI,CAAC,QAAQ;AACT,UAAM,IAAI,MAAM,sGAAsG;AAAA,EAC1H;AAEA,QAAM,cAAc,CAAC,SAAiB,WAAgC,CAAC,MAAM;AACzE,WAAO,IAAI,yBAAyB,SAAS;AAAA,MACzC;AAAA,MACA;AAAA,MACA,SAAS,QAAQ;AAAA,MACjB,GAAG;AAAA,IACP,CAAC;AAAA,EACL;AAEA,QAAM,WAAW,SAAU,SAAiB,UAAgC;AACxE,WAAO,YAAY,SAAS,QAAQ;AAAA,EACxC;AAEA,WAAS,OAAO;AAEhB,SAAO;AACX;AAeO,IAAM,UAA2B,SAAU,SAAiB,UAAgC;AAC/F,QAAM,SAAS,QAAQ,IAAI;AAC3B,MAAI,CAAC,QAAQ;AACT,UAAM,IAAI,MAAM,2GAA2G;AAAA,EAC/H;AACA,SAAO,IAAI,yBAAyB,SAAS;AAAA,IACzC;AAAA,IACA,SAAS;AAAA,IACT,GAAG;AAAA,EACP,CAAC;AACL;AAEA,QAAQ,OAAO;","names":[]}

package/package.json

@@ -0,0 +1,60 @@
+{
+    "$schema": "https://json.schemastore.org/package.json",
+    "name": "@cencori/ai-sdk",
+    "version": "0.2.1",
+    "description": "Cencori AI SDK - The infrastructure layer for AI applications. Works with Vercel AI SDK, TanStack AI, and more.",
+    "main": "dist/index.js",
+    "module": "dist/index.mjs",
+    "types": "dist/index.d.ts",
+    "exports": {
+        ".": {
+            "types": "./dist/index.d.ts",
+            "import": "./dist/index.mjs",
+            "require": "./dist/index.js"
+        }
+    },
+    "files": [
+        "dist",
+        "README.md"
+    ],
+    "keywords": [
+        "ai",
+        "cencori",
+        "ai-sdk",
+        "vercel",
+        "tanstack",
+        "provider",
+        "streaming",
+        "llm",
+        "openai",
+        "anthropic",
+        "gemini",
+        "infrastructure",
+        "security",
+        "observability"
+    ],
+    "author": "FohnAI",
+    "license": "MIT",
+    "repository": {
+        "type": "git",
+        "url": "git+https://github.com/cencori/cencori.git"
+    },
+    "homepage": "https://cencori.com",
+    "bugs": {
+        "url": "https://github.com/cencori/cencori/issues"
+    },
+    "scripts": {
+        "build": "tsup",
+        "dev": "tsup --watch",
+        "prepublishOnly": "npm run build"
+    },
+    "peerDependencies": {
+        "@ai-sdk/provider": ">=1.0.0"
+    },
+    "devDependencies": {
+        "@ai-sdk/provider": "^3.0.0",
+        "@ai-sdk/provider-utils": "^3.0.0",
+        "tsup": "^8.0.0",
+        "typescript": "^5.3.0"
+    }
+}