Skip to content
Source
Sign inSign up
TanStack/ai
Open History to see the last commit

openai.md

670 lines18.8 KB

The OpenAI adapter provides access to OpenAI's models, including GPT-4o, GPT-5, image generation (DALL-E), text-to-speech (TTS), and audio transcription (Whisper).

Installation

npm install @tanstack/ai-openai

Basic Usage

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Hello!" }],
});

Chat Completions API

@tanstack/ai-openai ships two text adapters that hit different OpenAI endpoints. openaiText (default) calls the Responses API (/v1/responses). openaiChatCompletions calls the older Chat Completions API (/v1/chat/completions).

Pick whichever fits your wire format and feature needs:

openaiText (Responses)openaiChatCompletions (Chat Completions)
Endpoint/v1/responses/v1/chat/completions
Reasoning summariesYes — set modelOptions.reasoning.summary: 'auto' to surface reasoning text via REASONING_* eventsNo — reasoning tokens are still consumed but cannot be exposed
Wire-format compatibilityOpenAI-onlyMatches the older de-facto industry shape (Grok, Groq, OpenRouter, many local model servers)
Structured output streamingtext.format: { type: 'json_schema', strict: true } + stream: trueresponse_format: { type: 'json_schema', strict: true } + stream: true

Use openaiText when you want reasoning-summary streaming or OpenAI-specific Responses features. Use openaiChatCompletions when you're migrating off a Chat-Completions-style provider, share request-building code with other Chat-Completions adapters in your stack, or want the more battle-tested wire format.

import { chat } from "@tanstack/ai";
import { openaiChatCompletions } from "@tanstack/ai-openai";

const stream = chat({
  adapter: openaiChatCompletions("gpt-5.2"),
  messages: [{ role: "user", content: "Hello!" }],
});

With an explicit API key:

import { chat } from "@tanstack/ai";
import { createOpenaiChatCompletions } from "@tanstack/ai-openai";

const adapter = createOpenaiChatCompletions("gpt-5.2", {
  apiKey: process.env.OPENAI_API_KEY!,
  // organization, baseURL, headers — all optional
});

const stream = chat({
  adapter,
  messages: [{ role: "user", content: "Hello!" }],
});

Both adapters work identically with Structured Outputs — including stream: true — and accept the same modelOptions (temperature, top_p, max_tokens, stop, …). The reasoning section below applies to openaiText; openaiChatCompletions accepts modelOptions.reasoning.effort but cannot stream summary text.

Basic Usage - Custom API Key

import { chat } from "@tanstack/ai";
import { createOpenaiChat } from "@tanstack/ai-openai";

const adapter = createOpenaiChat(process.env.OPENAI_API_KEY!, {
  // ... your config options
});

const stream = chat({
  adapter: adapter("gpt-5.2"),
  messages: [{ role: "user", content: "Hello!" }],
});

Configuration

import { createOpenaiChat, type OpenAIChatConfig } from "@tanstack/ai-openai";

const config: Omit<OpenAIChatConfig, 'apiKey'> = {
  organization: "org-...", // Optional
  baseURL: "https://api.openai.com/v1", // Optional, for custom endpoints
};

const adapter = createOpenaiChat(process.env.OPENAI_API_KEY!, config);

Example: Chat Completion

import { chat, toServerSentEventsResponse } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";

export async function POST(request: Request) {
  const { messages } = await request.json();

  const stream = chat({
    adapter: openaiText("gpt-5.2"),
    messages,
  });

  return toServerSentEventsResponse(stream);
}

Example: With Tools

import { chat, toolDefinition } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { z } from "zod";

const getWeatherDef = toolDefinition({
  name: "get_weather",
  description: "Get the current weather",
  inputSchema: z.object({
    location: z.string(),
  }),
});

const getWeather = getWeatherDef.server(async ({ location }) => {
  // Fetch weather data
  return { temperature: 72, conditions: "sunny" };
});

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages,
  tools: [getWeather],
});

Model Options

OpenAI supports various provider-specific options:

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages,
  modelOptions: {
    temperature: 0.7,
    max_tokens: 1000,
    top_p: 0.9,
    frequency_penalty: 0.5,
    presence_penalty: 0.5,
    stop: ["END"],
  },
});

Reasoning

Enable reasoning for models that support it (e.g., GPT-5, O3). This allows the model to show its reasoning process, which is streamed as thinking chunks:

modelOptions: {
  reasoning: {
    effort: "medium", // "none" | "minimal" | "low" | "medium" | "high"
    summary: "detailed", // "auto" | "detailed" (optional)
  },
}

When reasoning is enabled, the model's reasoning process is streamed separately from the response text and appears as a collapsible thinking section in the UI.

Summarization

Summarize long text content:

import { summarize } from "@tanstack/ai";
import { openaiSummarize } from "@tanstack/ai-openai";

const result = await summarize({
  adapter: openaiSummarize("gpt-5-mini"),
  text: "Your long text to summarize...",
  maxLength: 100,
  style: "concise", // "concise" | "bullet-points" | "paragraph"
});

console.log(result.summary);

Image Generation

Generate images with DALL-E:

import { generateImage } from "@tanstack/ai";
import { openaiImage } from "@tanstack/ai-openai";

const result = await generateImage({
  adapter: openaiImage("gpt-image-1"),
  prompt: "A futuristic cityscape at sunset",
  numberOfImages: 1,
  size: "1024x1024",
});

console.log(result.images);

Image Model Options

const result = await generateImage({
  adapter: openaiImage("gpt-image-1"),
  prompt: "...",
  modelOptions: {
    quality: "hd", // "standard" | "hd"
    style: "natural", // "natural" | "vivid"
  },
});

Text-to-Speech

Generate speech from text:

import { generateSpeech } from "@tanstack/ai";
import { openaiTTS } from "@tanstack/ai-openai";

const result = await generateSpeech({
  adapter: openaiTTS("tts-1"),
  text: "Hello, welcome to TanStack AI!",
  voice: "alloy",
  format: "mp3",
});

// result.audio contains base64-encoded audio
console.log(result.format); // "mp3"

TTS Voices

Available voices: alloy, echo, fable, onyx, nova, shimmer, ash, ballad, coral, sage, verse

TTS Model Options

const result = await generateSpeech({
  adapter: openaiTTS("tts-1-hd"),
  text: "High quality speech",
  modelOptions: {
    speed: 1.0, // 0.25 to 4.0
  },
});

Transcription

Transcribe audio to text:

import { generateTranscription } from "@tanstack/ai";
import { openaiTranscription } from "@tanstack/ai-openai";

const result = await generateTranscription({
  adapter: openaiTranscription("whisper-1"),
  audio: audioFile, // File object or base64 string
  language: "en",
});

console.log(result.text); // Transcribed text

Transcription Model Options

const result = await generateTranscription({
  adapter: openaiTranscription("whisper-1"),
  audio: audioFile,
  modelOptions: {
    response_format: "verbose_json", // Get timestamps
    temperature: 0,
    prompt: "Technical terms: API, SDK",
  },
});

// Access segments with timestamps
console.log(result.segments);

Environment Variables

Set your API key in environment variables:

OPENAI_API_KEY=sk-...

API Reference

openaiText(config?)

Creates an OpenAI chat adapter using environment variables.

Returns: An OpenAI chat adapter instance.

createOpenaiChat(apiKey, config?)

Creates an OpenAI chat adapter with an explicit API key.

Parameters:

  • apiKey - Your OpenAI API key
  • config.organization? - Organization ID (optional)
  • config.baseURL? - Custom base URL (optional)

Returns: An OpenAI chat adapter instance.

openaiChatCompletions(model)

Creates an OpenAI chat adapter that targets /v1/chat/completions instead of the Responses API. See Chat Completions API for when to use this over openaiText.

Returns: An OpenAI chat adapter instance using the Chat Completions wire format.

createOpenaiChatCompletions(model, config)

Creates an OpenAI chat-completions adapter with an explicit API key.

Parameters:

  • model - OpenAI model id (e.g. "gpt-5.2", "gpt-4o-mini")
  • config.apiKey - Your OpenAI API key
  • config.organization? - Organization ID (optional)
  • config.baseURL? - Custom base URL (optional)
  • config.headers? - Additional headers (optional)

Returns: An OpenAI chat adapter instance using the Chat Completions wire format.

openaiSummarize(config?)

Creates an OpenAI summarization adapter using environment variables.

Returns: An OpenAI summarize adapter instance.

createOpenaiSummarize(apiKey, config?)

Creates an OpenAI summarization adapter with an explicit API key.

Returns: An OpenAI summarize adapter instance.

openaiImage(config?)

Creates an OpenAI image generation adapter using environment variables.

Returns: An OpenAI image adapter instance.

createOpenaiImage(apiKey, config?)

Creates an OpenAI image generation adapter with an explicit API key.

Returns: An OpenAI image adapter instance.

openaiTTS(config?)

Creates an OpenAI TTS adapter using environment variables.

Returns: An OpenAI TTS adapter instance.

createOpenaiTTS(apiKey, config?)

Creates an OpenAI TTS adapter with an explicit API key.

Returns: An OpenAI TTS adapter instance.

openaiTranscription(config?)

Creates an OpenAI transcription adapter using environment variables.

Returns: An OpenAI transcription adapter instance.

createOpenaiTranscription(apiKey, config?)

Creates an OpenAI transcription adapter with an explicit API key.

Returns: An OpenAI transcription adapter instance.

Next Steps

Provider Tools

OpenAI exposes several native tools beyond user-defined function calls. Import them from @tanstack/ai-openai/tools and pass them into chat({ tools: [...] }).

For the full concept, a comparison matrix, and type-gating details, see Provider Tools.

webSearchTool

Enables the model to run a web search and return grounded results with citations. Pass a WebSearchToolConfig object (typed from the OpenAI SDK) to configure the tool.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { webSearchTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "What's new in AI this week?" }],
  tools: [webSearchTool({ type: "web_search" })],
});

Supported models: GPT-4o, GPT-5, and Responses API-capable models. See Provider Tools.

webSearchPreviewTool

The preview variant of web search with additional options for controlling search context size and user location. Use this when you want fine-grained control over the search context sent to the model.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { webSearchPreviewTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Latest news about TypeScript" }],
  tools: [
    webSearchPreviewTool({
      type: "web_search_preview_2025_03_11",
      search_context_size: "high",
    }),
  ],
});

Supported models: GPT-4o and above. See Provider Tools.

fileSearchTool

Searches OpenAI vector stores that you have pre-populated, letting the model retrieve relevant document chunks. Provide the vector_store_ids to search and optionally limit results with max_num_results (1–50).

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { fileSearchTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "What does the handbook say about PTO?" }],
  tools: [
    fileSearchTool({
      type: "file_search",
      vector_store_ids: ["vs_abc123"],
      max_num_results: 5,
    }),
  ],
});

Supported models: GPT-4o and above. See Provider Tools.

imageGenerationTool

Allows the model to generate images inline during a conversation using DALL-E/GPT-Image. Pass quality, size, and style options via the config object.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { imageGenerationTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Draw a logo for my app" }],
  tools: [
    imageGenerationTool({
      quality: "high",
      size: "1024x1024",
    }),
  ],
});

Supported models: GPT-5 and GPT-Image-capable models. See Provider Tools.

codeInterpreterTool

Gives the model a sandboxed Python execution environment. The container field configures the execution environment; pass the full CodeInterpreterToolConfig object.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { codeInterpreterTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Analyse this CSV and plot a chart" }],
  tools: [
    codeInterpreterTool({ type: "code_interpreter", container: { type: "auto" } }),
  ],
});

Supported models: GPT-4o and above. See Provider Tools.

mcpTool

Connects the model to a remote MCP (Model Context Protocol) server, exposing all its capabilities as callable tools. Provide either server_url or connector_id — not both.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { mcpTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "List my GitHub issues" }],
  tools: [
    mcpTool({
      server_url: "https://mcp.example.com",
      server_label: "github",
    }),
  ],
});

Supported models: GPT-4o and above. See Provider Tools.

computerUseTool

Lets the model observe a virtual desktop via screenshots and interact with it using keyboard and mouse events. Provide the display dimensions and the execution environment type.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { computerUseTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("computer-use-preview"),
  messages: [{ role: "user", content: "Open Chrome and navigate to example.com" }],
  tools: [
    computerUseTool({
      type: "computer_use_preview",
      display_width: 1024,
      display_height: 768,
      environment: "browser",
    }),
  ],
});

Supported models: computer-use-preview. See Provider Tools.

localShellTool

Provides the model with a local shell for executing system commands. Takes no arguments — the tool is enabled simply by including it in the tools array.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { localShellTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Run the test suite and summarise failures" }],
  tools: [localShellTool()],
});

Supported models: GPT-5.x and other agent-capable models. See Provider Tools.

shellTool

A function-style shell tool that exposes shell execution as a structured function call. Takes no arguments.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { shellTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Count lines in all JS files" }],
  tools: [shellTool()],
});

Supported models: GPT-5.x and other agent-capable models. See Provider Tools.

applyPatchTool

Lets the model apply unified-diff patches to modify files directly. Takes no arguments — include it in the tools array to enable patch application.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { applyPatchTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Fix the import paths in src/index.ts" }],
  tools: [applyPatchTool()],
});

Supported models: GPT-5.x and other agent-capable models. See Provider Tools.

customTool

Defines a custom Responses API tool with an explicit name, description, and format. Use this when none of the structured tool types fits your use case. Unlike branded provider tools, customTool returns a plain Tool and is accepted by any chat model.

import { chat } from "@tanstack/ai";
import { openaiText } from "@tanstack/ai-openai";
import { customTool } from "@tanstack/ai-openai/tools";

const stream = chat({
  adapter: openaiText("gpt-5.2"),
  messages: [{ role: "user", content: "Look up order #1234" }],
  tools: [
    customTool({
      type: "custom",
      name: "lookup_order",
      description: "Look up the status of a customer order by order ID",
    }),
  ],
});

Supported models: all Responses API models. See Provider Tools.