text-to-text

Text-to-Text: The core generative AI capability: It ingests a text prompt or document and outputs a transformed, coherent text result (summaries, translations, code).

This technology is the engine behind modern Large Language Models (LLMs), fundamentally a text-in, text-out system. It leverages deep learning and transformer architectures to process vast datasets, allowing for complex linguistic tasks. Specific applications include real-time content summarization (reducing 10,000 words to a 500-word brief), machine translation (e.g., Google Translate’s 100+ languages), and advanced generative tasks (drafting emails, writing code). The goal is high-fidelity text transformation, enabling automated content creation and immediate knowledge synthesis across enterprise platforms.

https://cloud.google.com/natural-language