Automated Text Sentence Extraction and Fact-Checking Workflow

This n8n workflow is designed to process large blocks of text, split it into individual sentences, and perform fact-checking using advanced language models. It begins with a manual trigger for testing purposes, where input text is provided or fetched from other workflows. A custom Code node efficiently splits the input into sentences, carefully preserving dates and list items.

The sentences are then split out for independent analysis. The workflow employs n8n’s language model integrations, including Ollama and Langchain, to process the sentences for fact-checking and summarization. Specific nodes utilize specialized models, such as ‘bespoke-minicheck’, for verifying factual accuracy, and others generate summaries or perform content validation.

The workflow allows for integration with other workflows, consolidates analysis results, and provides options for filtering and aggregating insights. This setup is useful for content verification, automated review processes, or enhancing data accuracy in publishing and data management scenarios. Overall, this workflow exemplifies how n8n can automate complex language understanding and validation tasks with AI models.