SummaryXL

SummaryXL is an Apify actor that automatically generates comprehensive AI summaries from any dataset. It processes your data as plain text and uses GPT-4o-mini to produce a detailed summary, helping you quickly understand large volumes of information without manual analysis.

For a complete tutorial on using SummaryXL to analyse Reddit comments in bulk, click here.

Key Features

• AI-Powered Summaries: Uses GPT-4o-mini to generate coherent, long-form summaries.
• Customizable Length: Set your summary length between 100 and 2000 words.
• Topic Focus: Optionally provide a focusTopic to steer the summary towards specific themes.
• Cost Tracking: Output includes detailed token usage and estimated API costs.
• Flexible Output: Receive summaries as a dataset record, a formatted PDF, or via optional email delivery.
• Seamless Integration: Works as an integration with any other Apify actor that outputs a dataset.

How to Use

As an Integration (Recommended)

1. Add the SummaryXL integration to your primary actor (e.g., a Reddit or Twitter scraper).
2. In the integration's datasetId field, enter {{resource.defaultDatasetId}} to automatically pass the dataset.
3. Configure your openaiApiKey and any other parameters.
4. Run your main actor. When it finishes, SummaryXL will process the dataset and output the summary.

For Direct Testing

You can run SummaryXL directly by providing a specific datasetId and your openaiApiKey in the actor input.

Input Parameters

Output Formats

The actor provides the summary in multiple formats:

1. Dataset Record: A JSON object containing the summary text, token usage, and cost breakdown.
   json { "summary": "The AI-generated summary text...", "tokenUsage": { "promptTokens": 1234, "completionTokens": 567, "totalTokens": 1801 }, "costs": { "promptCost": 0.001234, "completionCost": 0.000567, "totalCost": 0.001801 } }

2. PDF File: A formatted PDF is saved to the key-value store under the key OUTPUT, accessible via the Apify console.

3. Email (Optional): If EMAIL_SUPPORT is enabled and a targetEmail is provided, the summary is sent in HTML and plain text formats.

Configuration & Environment

Control additional behavior via these environment variables:
* EMAIL_SUPPORT: Enable/disable email functionality (default: true).
* MAX_PARALLEL_REQUESTS: Control parallel API requests (default: 10).
* PROMPT_TOKEN_COST: Cost per 1M prompt tokens for tracking (default: 0.150).
* COMPLETION_TOKEN_COST: Cost per 1M completion tokens for tracking (default: 0.075).

Example Input Configurations

Basic Integration:

{
  "openaiApiKey": "your-openai-api-key",
  "summaryLength": 1500,
  "focusTopic": "pricing trends"
}

Direct Test Run:

{
  "datasetId": "your-dataset-id",
  "openaiApiKey": "your-openai-api-key",
  "summaryLength": 1000
}

With Email Delivery:

{
  "openaiApiKey": "your-openai-api-key",
  "targetEmail": "user@example.com",
  "focusTopic": "customer feedback"
}