Palmyra Vision

Palmyra Vision is part of Writer’s Palmyra model family, extended to handle visual inputs along with text prompts. You feed it an image—anything from a chart, screenshot, product photo, or scanned document—and it produces actionable text: extracting text (even handwritten), summarizing or interpreting charts/graphs/infographics, classifying items or attributes, answering questions posed about the image, and generating descriptions.

It performs strongly on benchmarks like VQAv2 (visual question answering), and its design is enterprise-friendly: it works with structured output (so you can parse or automate downstream), it operates via API and tools in Writer’s platform, supports image- and graph-driven workflows, and is priced per image, second of video, or per million text tokens depending on input type. Use cases include product description generation from photos, compliance checks (detecting whether images meet brand/regulatory rules), transforming reports with charts into text summaries, digitizing handwritten notes, and improving accessibility by generating alt text for images.