AI Gateway now supports TypeSafe clients and an HTTP API for Jev
Vercel's AI Gateway now routes TypeSafe's Jev probabilistic model through a unified API surface—typed evals without the parse-and-retry tax.
Why it matters
Practical integration news: practitioners building with structured evaluations (agents, workflows, durable tasks) can now consolidate model calls and observability through AI Gateway. Reduces boilerplate for teams already on Vercel infrastructure.
The key facts
11 to knowJev (TypeSafe probabilistic model) now accessible via AI Gateway in three paths: TypeSafe client, HTTP API, AI SDK
Three question types: boolean (0-1 probability), choices (pick from named set), score (rate on custom scale)
Unified billing and observability across all three integration paths
Native integration with Eve framework (durable agents with sandboxed compute and built-in evals)
No breaking changes to existing TypeSafe client calls—URL and API key swap only
Published: September 21, 2026
Jev now accessible via TypeSafe client, HTTP API, and AI SDK
Three question types: boolean (0-1 probability), choices (multi-option picker), score (scale rating)
Requests unified in AI Gateway usage and observability alongside other model calls
Integration with eve framework for agent evaluation and tool approval automation
No SDK changes needed — base URL and API key swap only
Go to the source
Vercel Blogvercel.com
Publisher excerpt: You can now call through AI Gateway using an existing TypeSafe client or the HTTP API, in addition to the AI SDK.Jev from TypeSafe AI is a probabilistic decision model for software. State goes in, and typed answers come out with probabilities attached, so there's no generated text to parse.…