# High memory usage when processing many images

## Issue

When using `generateText` or `streamText` with many images (e.g., in a loop or batch processing), you may notice:

- Memory usage grows continuously and doesn't decrease
- Application eventually runs out of memory
- Memory is not reclaimed even after garbage collection

This is especially noticeable when using `experimental_download` to process images from URLs, or when sending base64-encoded images in prompts.

## Background

By default, the AI SDK includes the full request and response bodies in the step results. When processing images, the request body contains the base64-encoded image data, which can be very large (a single image can be 1MB+ when base64 encoded). If you process many images and keep references to the results, this data accumulates in memory.

For example, processing 100 images of 500KB each would include ~50MB+ of request body data in memory.

## Solution

Use the `experimental_include` option to disable inclusion of request and/or response bodies:

```ts
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';

const result = await generateText({
  model: openai('gpt-4o'),
  messages: [
    {
      role: 'user',
      content: [
        { type: 'text', text: 'Describe this image' },
        { type: 'image', image: imageUrl },
      ],
    },
  ],
  // Disable inclusion of request body to reduce memory usage
  experimental_include: {
    requestBody: false,
    responseBody: false,
  },
});
```

### Options

The `experimental_include` option accepts:

- `requestBody`: Set to `false` to exclude the request body from step results. This is where base64-encoded images are stored. Default: `true`. Available in both `generateText` and `streamText`.
- `responseBody`: Set to `false` to exclude the response body from step results. Default: `true`. Only available in `generateText`.

### When to use

- **Batch processing images**: When processing many images in a loop
- **Long-running agents**: When an agent may process many images over its lifetime
- **Memory-constrained environments**: When running in environments with limited memory

### Trade-offs

When you disable body inclusion:

- You won't have access to `result.request.body` or `result.response.body`
- Debugging may be harder since you can't inspect the raw request/response
- If you need the bodies for logging or debugging, consider extracting the data you need before the next iteration

## Example: Processing multiple images

```ts
import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';

const imageUrls = [
  /* array of image URLs */
];
const results = [];

for (const imageUrl of imageUrls) {
  const result = await generateText({
    model: openai('gpt-4o'),
    messages: [
      {
        role: 'user',
        content: [
          { type: 'text', text: 'Describe this image' },
          { type: 'image', image: imageUrl },
        ],
      },
    ],
    experimental_include: {
      requestBody: false,
    },
  });

  // Only store the text result, not the full result object
  results.push(result.text);
}
```

## Learn more

- [`generateText` API Reference](/docs/reference/ai-sdk-core/generate-text)
- [`streamText` API Reference](/docs/reference/ai-sdk-core/stream-text)


## Navigation

- [Azure OpenAI Slow to Stream](/docs/troubleshooting/azure-stream-slow)
- [Server Actions in Client Components](/docs/troubleshooting/server-actions-in-client-components)
- [useChat/useCompletion stream output contains 0:... instead of text](/docs/troubleshooting/strange-stream-output)
- [Streamable UI Errors](/docs/troubleshooting/streamable-ui-errors)
- [Tool Invocation Missing Result Error](/docs/troubleshooting/tool-invocation-missing-result)
- [Streaming Not Working When Deployed](/docs/troubleshooting/streaming-not-working-when-deployed)
- [Streaming Not Working When Proxied](/docs/troubleshooting/streaming-not-working-when-proxied)
- [Getting Timeouts When Deploying on Vercel](/docs/troubleshooting/timeout-on-vercel)
- [Unclosed Streams](/docs/troubleshooting/unclosed-streams)
- [useChat Failed to Parse Stream](/docs/troubleshooting/use-chat-failed-to-parse-stream)
- [Server Action Plain Objects Error](/docs/troubleshooting/client-stream-error)
- [useChat No Response](/docs/troubleshooting/use-chat-tools-no-response)
- [Custom headers, body, and credentials not working with useChat](/docs/troubleshooting/use-chat-custom-request-options)
- [TypeScript performance issues with Zod and AI SDK 5](/docs/troubleshooting/typescript-performance-zod)
- [useChat "An error occurred"](/docs/troubleshooting/use-chat-an-error-occurred)
- [Repeated assistant messages in useChat](/docs/troubleshooting/repeated-assistant-messages)
- [onFinish not called when stream is aborted](/docs/troubleshooting/stream-abort-handling)
- [Tool calling with structured outputs](/docs/troubleshooting/tool-calling-with-structured-outputs)
- [Abort breaks resumable streams](/docs/troubleshooting/abort-breaks-resumable-streams)
- [streamText fails silently](/docs/troubleshooting/stream-text-not-working)
- [Streaming Status Shows But No Text Appears](/docs/troubleshooting/streaming-status-delay)
- [Stale body values with useChat](/docs/troubleshooting/use-chat-stale-body-data)
- [Type Error with onToolCall](/docs/troubleshooting/ontoolcall-type-narrowing)
- [Unsupported model version error](/docs/troubleshooting/unsupported-model-version)
- [Object generation failed with OpenAI](/docs/troubleshooting/no-object-generated-content-filter)
- [Missing Tool Results Error](/docs/troubleshooting/missing-tool-results-error)
- [Model is not assignable to type "LanguageModelV1"](/docs/troubleshooting/model-is-not-assignable-to-type)
- [TypeScript error "Cannot find namespace 'JSX'"](/docs/troubleshooting/typescript-cannot-find-namespace-jsx)
- [React error "Maximum update depth exceeded"](/docs/troubleshooting/react-maximum-update-depth-exceeded)
- [Jest: cannot find module '@ai-sdk/rsc'](/docs/troubleshooting/jest-cannot-find-module-ai-rsc)
- [High memory usage when processing many images](/docs/troubleshooting/high-memory-usage-with-images)


[Full Sitemap](/sitemap.md)