Run open-weight OCR

VLM Run provides a unified API for deploying open-weight vision language models, OCR models, and ViT-based vision models in production. Built to simplify the operational complexity of serving these models at scale, addressing inconsistencies in model quantization, serving parameters, and visual accuracy across providers.

The service standardizes access to multiple vision model architectures under a single OpenAI-compatible interface, eliminating common production footguns like inconsistent model quantization and serving parameter variations.