Dataset2GPT Dataset2GPT reads any dataset on Apify and uses GPT-based AI to perform a variety of text-processing tasks—such as summarization, analysis, classification, or transformation. Whether you want a comprehensive 2,000-word synopsis, key insights, sentiment analysis, or custom NLP transformations, Dataset2GPT can handle it. > For a complete tutorial on how to integrate Dataset2GPT to analyse Reddit comments in bulk, please see our tutorial using our gpt-powered summary actor here. ## Features - GPT-Powered Processing using GPT-4o-mini for best price/performance ratio - Flexible Output Length: Control how long or detailed the result should be (100–2,000 words or more) - Task Focus: Provide a specific topic or angle for the AI to focus on (e.g., “marketing analysis,” “customer feedback,” “sentiment extraction”) - Token Usage & Cost Tracking: Monitor how many tokens the AI uses and the associated cost - PDF Export with Markdown formatting for easy sharing - Optional Email Delivery of processed results ## How It Works 1. Input: Your Apify actor collects or generates data in a dataset. 2. Processing: Dataset2GPT reads the dataset content as plain text, then uses GPT to perform your desired task. 3. AI Output: The AI generates the requested output (e.g., summary, analysis, transformation). 4. Delivery: Retrieve the output in multiple formats (dataset record, PDF, email). ## Benefits - Rapid Insights: Quickly extract meaning or structure from large datasets. - Plain Text Processing: No complex parsing required—just feed text. - Seamless Integration: Works with any Apify actor that outputs to a dataset. - Automation: Set up once and let Dataset2GPT handle the heavy lifting. ## Perfect For - Market Researchers - Social Media Managers - Data Scientists - Business Intelligence Teams - Content Analysts - Anyone working with scraped or collected data ## Input Parameters | Parameter | Type | Required | Default | Description | |--------------|---------|----------|---------|-------------------------------------------------------------------------------------------------| | datasetId | string | No | - | Dataset ID for testing. When used as an integration, dataset ID is handled automatically | | openaiApiKey | string | Yes | - | Your OpenAI API key for GPT-based processing | | summaryLength| integer | No | 1000 | Desired length of output in words (100–2000). You can also use it for controlling output detail | | focusTopic | string | No | - | Specific topic or angle to guide the AI’s processing (e.g., “sentiment analysis” or “key points”)| | targetEmail | string | No | - | Email address to send the results to (requires EMAIL_SUPPORT=true) | ## Output Formats 1. Dataset Record json { "output": "The AI-generated text (e.g., summary, analysis, transformations)...", "tokenUsage": { "promptTokens": 1234, "completionTokens": 567, "totalTokens": 1801 }, "costs": { "promptCost": 0.001234, "completionCost": 0.000567, "totalCost": 0.001801 } } - The output field contains the GPT-processed text. - Token usage and cost breakdown are provided for transparency. 2. PDF Output - Stored in a key-value store under the OUTPUT key. - Includes Markdown formatting for better readability. - Downloadable via the Apify console or API. 3. Email Delivery (Optional) - Sends the GPT output in HTML and plain text formats. - Requires EMAIL_SUPPORT=true and a valid targetEmail. ## Getting Started 1. Add Dataset2GPT to Your Apify Actor: For example, if you have an actor scraping Reddit or Twitter, attach Dataset2GPT as an integration step. 2. Use the Default Dataset: In your input configuration, set "datasetId": "{{resource.defaultDatasetId}}" to automatically consume the scraped data. 3. Run Your Apify Actor: Once it finishes, Dataset2GPT will read the dataset and produce the requested AI output. 4. Retrieve Your Results: - Download from the dataset. - Grab the PDF from the key-value store. - Or check your inbox if you enabled email delivery. ## Usage Examples ### As an Integration (Recommended) json { "openaiApiKey": "your-openai-api-key", "summaryLength": 1500, "focusTopic": "product feedback analysis" } > Here, Dataset2GPT will generate a ~1500-word analysis focusing on product feedback. ### For Testing json { "datasetId": "your-dataset-id", "openaiApiKey": "your-openai-api-key", "summaryLength": 1000 } > Useful for local testing. Specify a known datasetId. ### With Email Delivery json { "openaiApiKey": "your-openai-api-key", "targetEmail": "user@example.com", "focusTopic": "sentiment extraction" } > Generates text focusing on sentiments and sends it to user@example.com. ## Environment Variables - EMAIL_SUPPORT: Enable/disable email functionality (default: true) - MAX_PARALLEL_REQUESTS: Controls parallel API requests (default: 10) - PROMPT_TOKEN_COST: Cost per 1M prompt tokens (default: 0.150) - COMPLETION_TOKEN_COST: Cost per 1M completion tokens (default: 0.075) --- With Dataset2GPT, turn any dataset into meaningful insights, summaries, transformations, or advanced NLP tasks—powered by GPT.