How I fed 60 million tokens to stealth/ox-alpha, or the story of how 127 nodes were 'vibe-coded' in n8n

The author shares the results of an engineering experiment involving the processing of 60 million tokens using the stealth/ox-alpha model. The article details the creation of a complex automation workflow in the n8n platform, which eventually grew to 127 nodes. Departing from a previously planned analysis of hybrid data extraction (LLM-tools + regex), the focus shifts to practical aspects of handling large datasets and workflow automation. The author notes that the decision was made spontaneously during development and invites readers to explore the technical implementation details of such a large-scale automation system. This article is of interest to professionals working with LLMs, business process automation, and the integration of complex data processing pipelines.