From the source
Interfaze runs its model-serving backend on Vercel, handling 50 million API requests a month for engineering teams in legal, healthcare and finance.
- The company builds models for deterministic work that general-purpose LLMs handle poorly, including document processing, OCR, object detection, web scraping, speech to text and classification.
- Founder Yoeven Khemlani says an LLM asked for an invoice total returns a number with no easy way to check its origin, so Interfaze's models return the number with a bounding box and a confidence score.
- The backend runs as serverless functions in the same codebase as the frontend, with Vercel Sandbox executing code the model writes and Fluid compute scaling with traffic.
- AI Gateway fails requests over to another provider when one throttles or goes down, and lets Interfaze compare providers serving the same open source model at different quality.
- The Workflow SDK runs lifecycle email sequences with durable execution, and preview deployments carry new model versions to customers before release.
- Interfaze recently released Interfaze 1 Lite, an open-weight version of its model that runs on a single GPU and is served through the same API.
