Writer’s latest models can generate text from images, including charts and graphs

Writer’s latest models can generate text from images, including charts and graphs

Writer, a San Francisco startup, is making waves in the realm of generative AI, distinguishing itself with practical business applications. The recent introduction of Palmyra-Vision, an enhancement to the Palmyra model, allows the generation of text from images, particularly graphs and charts. The strategic focus on multimodal content positions Writer uniquely, concentrating on text generation and insights derived from text-based outputs. May Habib, the CEO, emphasized the versatile use cases, ranging from e-commerce websites automatically updating content to compliance checking in pharmaceutical advertising.

Employing a multi-model approach, Writer achieves impressive accuracy (four nines) in determining image content and generating corresponding text. The applications span various sectors, including e-commerce, compliance, and interpretation of handwritten notes. Habib stressed the necessity of human review in the workflow, acknowledging the potential for model errors or hallucination. Writer, having secured $126 million in funding, is actively exploring partnerships with major cloud infrastructure platforms to facilitate scaling. The latest Palmyra release, featuring image-to-text capabilities, is set to impact diverse industries, providing a glimpse into the future of generative AI.